Original Paper
Caries Res 1999;33:93–100
Incidence and Progression of Approximal Caries
from 11 to 22 Years of Age in Sweden:
A Prospective Radiographic Study
I. Mejàre a C. Källestål b, c H. Stenlund b, d
a Department of Paediatric Dentistry, Eastman Dental Institute, Stockholm, and Departments of
b Epidemiology and Public Health, c Pedodontics and d Statistics, University of Umeå, Sweden
Key Words
Adolescence · Approximal caries · Caries rate · Cohort ·
Incidence · Progression · Radiography · Teenage
Abstract
Using annual bite-wing radiographs, the incidence
and progression of approximal caries (4d–7m) were
assessed longitudinally in teenagers and adolescents
whose treatment had been based on remineralizing
rather than restorative strategies. A closed cohort of 536
children initially was followed from 11 to 22 years of
age. The scoring system was: 0 = no visible radiolucen-
cy; 1–2 = radiolucency in the enamel up to the enamel-
dentin border; 3 = radiolucency with a broken enamel-
dentin border but with no obvious progression in the
dentin; 4 = radiolucency with obvious spread in the out-
er half of the dentin, and 5 = radiolucency in the inner
half of the dentin. Caries rates were estimated as the
number of new lesions/100 tooth surface-years, and the
Kaplan-Meier estimate was used to calculate the cumu-
lative survival time of each approximal surface. Three
events were used: the transitions from states 0 to 2, 2 to
4 and 3 to 4. The results showed a considerable varia-
tion between the surfaces in both caries rates and sur-
vival time. For all surfaces combined, the median caries
rate from state 0 to 2 was 3.9 new lesions/100 tooth sur-
face-years; from state 2 to 4, the rate was 5.4, and from
state 3 to 4 it was 20.3. Of the sound surfaces (state 0),
75% survived 6.3 years without reaching state 2. Given
쑕 1999 S.Karger AG, Basel
0008–6568/99/0332–0093 $17.50/0
Fax +41 61 306 12 34
E-Mail karger@karger.ch Accessible online at:
www.karger.com http://BioMedNet.com/karger
Received: May 12, 1998
Accepted after revision: July 6, 1998
state 2, 75% survived 4.8 years without reaching the out-
er half of the dentin (state 4), while given a lesion at the
enamel-dentin border (state 3), 75% survived 1.3 years
without doing the same. The median survival time of
lesions from state 3 to 4 was 3.1 years. The group
with DMFSappr c1 at the age of 11–12 years had a risk of
new approximal enamel lesions (state 0–2) that was
2.5 times greater than that of the group with
DMFSappr = 0–1.
For deciding when to restore an approximal lesion, two
of the main considerations are if and how rapidly the lesion
will progress if left unrestored. A slow rate of progression
through the enamel has been reported in a number of stud-
ies, justifying remineralizing procedures instead of restora-
tive treatment of such lesions [Pitts, 1983; Shwartz et al.,
1984a]. In recent years, moreover, a treatment strategy has
been reported that involves monitoring lesion progression
even beyond the enamel-dentin border [Holst et al., 1986;
Heidmann et al., 1987; Bille and Carstens, 1989; Lith et al.,
1995], though little is known about the rate of progression
in the outer part of the dentin, partly because the number of
lesions available for analysis is small and/or biased by le-
sions being restored so that the rate of progression if left un-
restored could not be determined [Shwartz et al., 1984a].
More therefore needs to be known about the progression of
approximal caries lesions through the enamel and into the
outer part of the dentin.
200.89.68.56 - 3/31/2015 4:58:38 PM
Ingegerd Mejàre
Universidad de Chile
Eastman Dental Institute
Dalagatan 11
Downloaded by:
S–11324 Stockholm (Sweden)
Tel. +46 08 7298936, Fax +46 08 310488, E-Mail ingegerd.mejare@ftv.sll.se

Fig. 1. The three events used to assess approximal caries rate and
progression: the transitions from states 0 to 2, 2 to 4 and 3 to 4.
Determining the rate and pattern of incidence and pro-
gression of caries on individual surfaces requires repeated
observations of a large group over a rather long period. In a
long-term study, the development of approximal lesions
through the enamel and into the outer part of the dentin was
monitored regularly in a cohort from the age of 11 to 22
years [Mejàre et al., 1998]. The patients attended a compre-
hensive Public Dental Health Service where the treatment
philosophy was characterized by a conservative approach,
implying remineralizing measures and monitoring the pro-
gression of approximal lesions rather than restoring them.
Based on these data, the aim of the present study was to as-
sess radiographically the incidence and progression of ap-
proximal caries through the enamel and into the outer part
of the dentin.
Materials and Methods
The original material, fully described elsewhere [Mejàre et al.,
1998], consisted of bite-wing radiographs from 536 children, 11–13
years old at baseline in 1985–1986. All children belonging to the
catchment area of the Västerhaninge clinic, in a southern suburb of
Stockholm, were included in the study. Mean DMFS scores at base-
line were for 11-year-olds 2.9, for 12-year-olds 3.6 and for 13-year-
olds 4.4. The cohort was closed and followed up to 21–22 years of
age. The treatment strategy was characterized by a preventive rather
than a restorative approach and apart from a clinically diagnosed cav-
ity, the general threshold for restorative treatment of an approximal
lesion was to be obvious progression in the outer half of the dentin,
that is score 4 (fig. 1). The last examination included 364 individuals.
As a rule, two posterior bite-wing radiographs were taken annually in
each side of the mouth with the aid of a filmholder (Kwik-bite, Troll-
hätteplast, Sweden). The mean number of sets of bite-wing radio-
graphs of each individual was 8.7.
To minimize errors from variations in interindividual treatment
decisions [Nuttall and Pitts, 1990; Kay et al., 1992], an advisory board
consisting of three dentists was constituted at the clinic. Before the in-
dividual dentist restored a proximal surface, the board was to be con-
sulted and a consensus achieved as to whether to restore or not.
94 Caries Res 1999;33:93–100
Table 1. Estimated age at eruption of per-
manent premolar and molar teeth
Tooth Age at eruption, years
upper lower
4 10.2 10.5
5 11.0 11.2
6 6.6 6.5
7 12.5 11.9
Modified after Kreiborg et al. [1994].
Classification and Coding of the Radiographs
The approximal surfaces were classified according to a scoring
system: 0 = no visible radiolucency; 1–2 = radiolucency in the enam-
el up to the enamel-dentin border; 3 = radiolucency with a broken
enamel-dentin border but with no obvious progression in the dentin; 4
= radiolucency with obvious spread in the outer half of the dentin, and
5 = radiolucency in the inner half of the dentin. The D component of
the DMF index included all surfaces with a radiolucency in the dentin,
that is score 3, 4 or 5. Restored and unreadable surfaces and extracted
teeth had separate codes. In coding the radiographs, reversals were
not allowed. Thus, small radiolucencies that disappeared on subse-
quent radiographs, perhaps due to remineralization, were coded as in-
tact surfaces throughout. Further, lesions that appeared to return to a
shallower state on later radiographs, perhaps because of a change in
the direction of the X-ray beam, were coded as being in the less deep
state throughout. In an assessment of this coding procedure, Shwartz
et al. [1984b] concluded that the effects on the estimated rate of pro-
gression would be minimal. The radiographs were analysed and cod-
ed by either of two of the authors (I.M. and C.K.). Reproducibility
was tested by means of κ values for the radiographic scores 0–5 [Co-
hen, 1960]. To calculate intra-examiner reproducibility, 10% of the ra-
diographs were analysed twice by each examiner and for the calcula-
tion of interexaminer agreement another 15% were analysed by both.
The interexaminer agreement had a κ value of 0.52 and the intra-ex-
aminer agreement gave κ values of 0.77 and 0.56.
Data Management and Statistical Analysis
The radiographic scores were regarded as carious states and three
events were used: the transitions from states 0 to 2, 2 to 4 and 3 to 4
(fig. 1). A transition from state 2 or 3 directly to state 5 or restoration
occurred only rarely. For each year, one set of radiographs/individual
was included in the statistical analysis. The primary purpose was to
assess the incidence and progression of caries for each approximal
surface (4d–7m). To do this, the duration of the states had to be deter-
mined, for example, the time a surface was ‘sound’ (state 0) before
progressing to state 2 or the time a lesion was in state 2 before
progressing to state 4. A starting point for each surface was represent-
ed by its eruption date, modified after Kreiborg et al. [1994] (table 1).
An alternative would be to use the date when the tooth reached
occlusion and contact with the neighbouring tooth. However, the
interval between eruption and occlusion is short, i.e. 10–14 weeks
[Hargreaves, 1958], and was considered to be within the margin of er-
ror. The use of eruption dates was considered to be more practical.
200.89.68.56 - 3/31/2015 4:58:38 PM
Mejàre/Källestål/Stenlund
Universidad de Chile
Downloaded by:
The eruption date of the distal surface of the first molar can be re-
garded as a special case, since contact with the neighbouring tooth is
not achieved until the second molar erupts; it was therefore set to be
equal to the eruption data of the neighbouring mesial surface of the
second molar.
A lesion that was in the next higher state at the following exami-
nation was assumed to have made this transition midway between the
two examinations.
Unseen values of the starting or end point of a state occurred when
a lesion was two or more states higher at the next examination or was
already in a carious state at baseline. Unseen values were estimated as
follows. A lesion that was two states higher at the next examination
was assumed to have progressed to the next higher state a quarter of
the way between the two examinations, to have remained in that state
until three quarters of the way between the examinations and then to
have progressed to the next state. Corresponding estimates were made
for lesions that were three states higher at the next examination. For
lesions that were in state 3 or 4 at baseline, the eruption date was used
to calculate the transition to previous states. For example, if a lesion
was scored 3 at baseline, the transition to score 2 was set at two thirds
of the interval between the eruption date and the baseline date. Alto-
gether, starting or end points had to be estimated for 8.7% of all the
events of state 2, 3 or 4. A little less than half of these estimates con-
cerned the date of the next higher state (from 1 to 3 or 2 to 4) and the
rest concerned the dates of two or more next higher states, for exam-
ple from 0 to 3 or 1 to 4. The mesial surfaces of the first molar ac-
counted for 41% of all these estimates.
Of the mesial surfaces of the first molar, 5.9% had already been re-
stored at the start. The rest, 2,012 surfaces, were included in the anal-
yses (table 2). Of these, 17.3% were already in one of the carious
states 2, 3, 4 or 5 at the start. 71 premolars and 1 molar were extract-
ed at baseline or during the study period and excluded from the anal-
yses.
Incidence was expressed as the crude incidence rate, referred to as
incidence density (ID) [Miettinen, 1976], calculated according to
Kleinbaum et al. [1982], using the formula:
ID(t0, t) = I/PT,
where I is the number of new lesions from states 0 to 2, 2 to 4 and 3 to
4, respectively, that occurred during the period t0, t (11–22 years of
age), and PT is the amount of risk time during the period expressed as
surface-years. ID is the number of new surfaces/surface-year, here-
after called caries rate.
In order to assess any dependence of baseline caries prevalence on
enamel and dentin caries rates, ratio comparisons (relative risk) of the
caries rates in two groups were calculated: DMFSappr c1 (= 28% of
the children) and DMFSappr = 0–1 at the age of 11–12. The 95% con-
fidence interval was used to test statistical significance. Any corre-
sponding dependence of caries rates on age was assessed by calculat-
ing the relative risk of caries rates from state 2 to 4 in two age groups:
11–16 years and 17–22 years of age. For states 2–4, ratio comparisons
of caries rates of lesions that had reached state 2 during the first age
interval were made with caries rates of lesions that reached state 2
during the second age interval. A corresponding procedure was used
for states 3–4.
Progression was assessed in terms of survival times from one state
to another. The Kaplan-Meier estimate [Kaplan and Meier, 1958] was
used and the cumulative survival time of each approximal surface for
the three events was calculated, that is the time during which a sur-
face/lesion remained in state 0–1 before reaching state 2, the time
Caries Incidence and Progression from the
Age of 11 to 22
from state 2 until the lesion reached state 4 or 5 or was restored and
the time from state 3 to state 4 or 5 or restoration.
For practical reasons, statistical analyses from states 0 to 1, 1 to 2,
1 to 3 and 1 to 4 were omitted. Surfaces that were still in state 0 or 1
at the end of the study or were withdrawn for other reasons constitut-
ed censored observations in the analysis of survival from state 0 to 2;
similarly, lesions that were still in state 2 or 3 at the end were censored
observations in the analysis of survival from state 2 to 4 (table 2),
while lesions that were still in state 3 at the end were censored obser-
vations in the analysis of survival from state 3 to 4. Because of the
large number of censored observations, the survival times are sum-
marized in percentiles instead of mean values. For example, the 75th
percentile in table 2 denotes the time until 25% of the surfaces had
reached state 2, i.e. 75% still surviving. The 50th percentile equals the
median value. A missing median value of the cumulative survival
times from one state to the next indicates that less than 50% of the sur-
faces had progressed to the next state. For practical reasons, all calcu-
lations were made under the assumption that the tooth surfaces were
independent. The data were processed in SPSS (Statistical Package
for Social Sciences) for Windows, version 7.5.
Results
There was no statistically significant difference between
males and females so their data were combined. For all sur-
faces combined, the median caries rates (new lesions/100
tooth surface-years) were 3.9 (range = 1.0–8.0) from state 0
to 2, 5.4 (range = 1.1–9.3) from state 2 to 4 and 20.3 (range =
4.6–28.6) from state 3 to 4. Survival times and caries rates of
each approximal surface from state 0 to 2 and state 2 to 4 are
presented in table 2. Of the ‘sound’ surfaces (state 0), 75%
survived 3.0–9.0 years (median value = 6.3 years) without
reaching state 2. Given state 2, 75% survived 3.4–6.6 years
(median value = 4.8 years) without reaching state 4. For the
right and left sides combined, the caries rates from state 0 to
state 2 were highest for the distal surface of the upper and
lower first molars and second premolars. Caries rates from
state 2 to 4 were highest for the mesial surface of the upper
second molars, the distal surface of the lower first molars and
the distal surface of the upper second premolars (fig. 2).
From state 0 to 2, 46d showed the shortest and 15m the
longest survival times. Their survival functions according to
the Kaplan-Meier estimate are illustrated in figure 3.
For lesions that had reached the enamel-dentin border
(state 3), the median cumulative survival time before they
reached state 4 ranged from 2.0 to 6.8 years (median
value = 3.1 years; fig. 4). Of state 3 lesions, 75% survived
1.3 years before reaching state 4. From state 3 to 4, the
highest caries rates were observed for the mesial surface of
the lower second molars, the distal surface of the upper sec-
ond premolars and the distal surface of the lower first mo-
lars (fig. 4).
200.89.68.56 - 3/31/2015 4:58:38 PM
Caries Res 1999;33:93–100 95
Universidad de Chile
Downloaded by:
Table 2. Survival times and caries rates (number of new lesions/100 tooth surface-years) in enamel and dentin of approximal surfaces (7m–4d)
from 11 to 22 years of age

Tooth States 0–2

surface
in analysis events censored excluded risk time survival time caries
n n n years years rate
n %
enamel
75th 50th (median)

17m 523 91 432 83 0 (13) 3,659 – – 2.5

16d 535 182 353 67 0 (1) 3,139 4.8 – 5.8
16m 509 229 280 55 26 (1) 5,733 8.3 – 4.0
15d 513 225 288 56 3 (20) 3,790 4.8 10.8 5.9
15m 514 133 381 74 2 (20) 4,261 9.0 – 3.1
14d 521 144 377 73 2 (13) 4,592 8.9 – 3.1
27m 521 110 411 79 0 (15) 3,427 8.8 – 3.2
26d 531 188 343 65 1 (4) 3,147 4.8 – 5.9
26m 495 224 271 55 38 (3) 5,663 7.7 15.6 4.0
25d 510 203 307 60 2 (24) 3,867 6.3 – 5.3
25m 512 142 369 73 2 (22) 4,199 8.2 – 3.4
24d 525 157 368 70 1 (10) 4,722 8.6 – 3.3
37m 527 218 309 59 3 (6) 3,605 5.3 – 6.0
36d 534 256 278 52 2 (0) 3,202 3.5 8.8 8.0
36m 507 212 295 58 29 (0) 5,634 8.3 – 3.8
35d 508 182 326 64 3 (25) 3,903 6.1 – 4.7
35m 509 55 454 89 0 (27) 4,464 – – 1.2
34d 1 526 56 470 90 1 (9) 4,904 – – 1.1
47m 529 177 352 67 0 (7) 3,864 6.3 – 4.6
46d 533 246 287 54 2 (1) 3,154 3.0 – 7.8
46m 501 206 295 59 34 (1) 5,795 8.8 – 3.6
45d 511 239 272 53 1 (24) 3,658 4.7 9.6 6.5
45m 1 507 48 459 91 2 (27) 4,576 – – 1.0
44d 1 525 84 441 84 0 (11) 4,939 – – 1.7

In analysis = Total number of analysed surfaces; events = lesions having reached state 2 and state 4, respectively; censored = lesions not hav-
ing reached state 2 and state 4, respectively; excluded = surfaces restored at baseline, figures in parentheses = extracted, unreadable or not erupt-
ed; risk time of states 0–2 = time interval between state 0 and 2 or the time that state 0 or 1 was observed; risk time of states 2–4 = time inter-
val between state 2 and 4 or the time that state 2 or 3 was observed; 75th, 50th = percentiles; – = no data due to too few events.
1 The proportion of censored lesions exceeds 90%.

For the majority of the surfaces, the risk of developing Discussion

new enamel lesions (from state 0 to 2) was significantly
higher for individuals with a DMFSappr c1 at the age of
11–12 years compared with those with a DMFSappr = 0–1
(table 3). Thus, in the group with a DMFSapp rc1, the risk
(median of the relative risk values) of new approximal
enamel lesions was 2.5 times higher than in those with a
DMFSappr = 0–1. The corresponding relative risk of dentin
caries (states 2–4) did not reach statistical significance for
any of the surfaces.
A comparison of caries rates from states 2 to 4 and 3 to 4
in the two age groups of 11–16 and 17–22 years showed
a statistically higher caries risk from state 2 to 4 in the
younger age group for three surfaces: 37m, 46d and 45d.
The drop-outs, of which the majority occurred at the end
of the study period, amounted to 32% of the original 536 in-
dividuals, which, considering the length of the study period,
must be regarded as acceptable. The reasons for the drop-
outs have been presented previously [Mejàre et al., 1998]
and were not considered to entail any notable bias. The de-
tailed radiographic recordings of the approximal surfaces,
with five depth categories defined and recorded, probably
explain the relatively low inter- and intra-examiner agree-
ments [Pliskin et al., 1984].
It was assumed that a change from one carious state to
the next occurred at the midpoint between two examina-
200.89.68.56 - 3/31/2015 4:58:38 PM

96 Caries Res 1999;33:93–100 Mejàre/Källestål/Stenlund

Universidad de Chile
Downloaded by:
States 2–4
in analysis events censored risk time survival time caries
n n years years rate
n %
dentin
75th 50th (median)

91 25 66 73 317 3.7 7.3 7.9

182 39 143 79 745 5.7 – 5.2
229 64 165 72 1,112 4.8 10.6 5.8
225 67 158 70 784 3.5 7.3 8.6
133 24 109 82 499 5.9 – 4.8
144 27 117 81 540 5.1 – 5.0
110 34 76 69 367 3.5 6.6 9.3
188 34 154 82 761 6.6 – 4.6
224 51 173 77 1,148 6.6 – 4.4
203 53 150 74 738 3.4 – 7.2
142 29 113 80 548 5.1 9.3 5.3
157 33 124 79 577 5.2 8.3 5.7
218 38 180 83 841 5.6 – 4.5
256 85 171 67 1,038 4.1 7.2 8.2
212 58 154 73 1,079 4.8 – 5.4
182 39 142 78 688 4.8 8.2 5.8
55 6 49 89 187 5.6 – 3.2
56 5 51 91 183 – – 2.8
177 44 133 75 649 4.6 7.0 6.8
246 86 160 65 1,033 4.1 6.7 8.3
206 64 142 69 988 3.8 – 6.5
239 44 195 82 979 6.3 – 4.5
48 3 45 94 144 – – 2.1
84 3 81 96 272 – – 1.1

tions. This assumption is not self-evident and it may be a
possible source of error, the magnitude of which is difficult
to assess. Median values of survival times from state 0 to 2
could not be estimated for most of the surfaces, because
50% of the surfaces in question had not reached state 2 by
the end of the study. This illustrates the slow progression
rate for approximal enamel lesions. It should be noted that
for some of the surfaces, particularly the lower premolars,
the number of censored lesions was high; these results must
therefore be interpreted with caution.
The mesial surface of the first molar was in an excep-
tional position because its eruption date had occurred sever-
al years before the start of the study. Of these surfaces, 23%
were either filled or in one of the analysed carious states at
the start. Unseen values from eruption date to the start of the
study were estimated for 17% of these surfaces, on the as-
sumption that the change from one state to the next oc-
curred midway between two examinations. The potential
bias from these estimates implies that both the caries rates
and the survival data for this surface should be interpreted
with caution.
The results show that 75% of the lesions survived c6
years (median value = 6.3 years) from state 0 to 2, which is
longer than was found in earlier studies in which the Ka-
plan-Meier estimate was used to estimate enamel caries
progression [Shwartz et al., 1984a; Lervik et al., 1990].
These authors estimated the mean time for a lesion to
progress through the entire enamel to be 4 to c7 years,
200.89.68.56 - 3/31/2015 4:58:38 PM

Caries Incidence and Progression from the Caries Res 1999;33:93–100 97

Universidad de Chile

Age of 11 to 22
Downloaded by:
Fig. 2. Enamel caries rates (states 0–2) ranked in ascending order of Fig. 3. Survival functions of 46d and 15m from state 0 to 2.
each jaw and dentin caries rates (states 2–4) of approximal surfaces
from 11 to 22 years of age. Right- and left-side surfaces are pooled.

Table 3. Significant relative caries rate risk

of approximal surfaces in two prevalence
groups: DMFSappr c1 versus DMFSappr = 0–1
at 11–12 years of age

Tooth Enamel caries (states 0–2)

surface
relative risk 95% confidence
interval

17m 2.8 1.35–5.90

16d 2.2 1.28–3.91
16m 2.7 1.41–5.30
14d 2.3 1.17–4.49
27m 3.1 1.44–6.85
26d 1.9 1.06–3.30
26m 2.6 1.30–5.19
25d 2.5 1.18–5.22
24d 2.3 1.20–4.50
Fig. 4. Approximal caries rates and median survival times in the 37m 3.7 1.71–8.11
dentin (from state 3 to 4) from 11 to 22 years of age. Right- and left- 36m 2.5 1.28–4.80
side surfaces are pooled and ranked in ascending order according to 35d 2.1 1.00–4.33
the caries rates of each jaw. Survival time of lower 4d is missing be- 35m 4.5 2.33–8.82
cause of too few observations. 34d 4.8 2.55–9.14
47m 2.2 1.04–4.52
46d 1.9 1.04–3.29
46m 2.7 1.40–5.23
depending on age and how censored lesions were handled.
Relative risk = Caries rate of group
The reason for this difference is most probably that also cen- DMFSappr c1/caries rate of group DMFSappr
sored surfaces were included in the present analysis. Thus, it = 0–1.
was assumed that the eruption data of each approximal sur-
face constituted its starting point for risk. The comparative-
ly long duration of risk that this entails affects the estimates
200.89.68.56 - 3/31/2015 4:58:38 PM

98 Caries Res 1999;33:93–100 Mejàre/Källestål/Stenlund

Universidad de Chile
Downloaded by:
of both caries rate and cumulative survival time from state 0
to 2. This way of estimating risk time is debatable; it was
chosen because a sound surface at the start of the study rep-
resents a retro-active period of risk. For example, a radio-
graphically sound mesial surface of the first molar at base-
line represented an existing risk period of 5–6 years.
While caries progression through the enamel has been
analysed in a number of studies [for a review see Pitts,
1983], data on the rate of progression in the dentin are
sparse. Emslie [1959] and Kohlemainen and Rytmömaa
[1977] reported a slow but unspecified progression in the
dentin. However, the populations in both these studies were
highly selected, the materials were small and the observa-
tion periods restricted to 2–3 years.
In the present study, progression in the dentin when the
lesion had radiographically reached the enamel-dentin bor-
der was notably faster than progression in the enamel. Thus,
the median caries rate was 3.8 times higher (20.3/5.4) from
state 3 to 4 than from state 2 to 4 and 5.2 times higher
(20.3/3.9) from state 3 to 4 than from state 0 to 2. On the
other hand, after 3.1 years, 50% of the lesions at the enam-
el-dentin border still did not show any obvious radiograph-
ically detectable progression in the dentin. This is consider-
ably slower than suggested by Pine and ten Bosch [1996],
who from an analysis of five studies [Haugejorden and
Slack, 1975; Gröndahl et al., 1977; Gröndahl and Hollen-
der, 1979; Bille and Carstens, 1989; Strang, 1990] conclud-
ed that 50% of approximal dentin lesions were filled or had
reached the pulp within 2 years and that the majority had
reached this stage by 3 years. However, ignoring restored
lesions and non-progressing lesions in the statistical analy-
sis introduces a bias towards the overestimation of progres-
sion [Shwartz et al., 1984c]. Furthermore, several of the
above-mentioned studies are at least 10 years old, so anoth-
er explanation for the different results could be that in the
meantime the rate of progression has continued to slow.
Caries rates and survival times varied considerably be-
tween the different surfaces, as shown in figures 3 and 4,
illustrating the site-specific development of approximal
caries. For all the events combined (states 0–2, 2–4 and
3–4), the distal surface of the upper premolar showed the
highest caries rates and the lowest survival times of all the
observed surfaces. The reason for this is not known but con-
sistent with the findings of Chestnut et al. [1996]. The high-
est caries rates from ‘sound’ to carious (from state 0 to 2)
were observed for distal surfaces of the upper and lower
first molars and premolars, indicating that the distal sur-
faces are more caries-prone than the mesial surfaces.
Carlos and Gittelsohn [1965] and Hujoel et al. [1994]
found that caries rate decreased with increasing age. A com-
Caries Incidence and Progression from the
Age of 11 to 22
parison of the younger and older age groups in the present
study showed statistically significant differences in caries
rate risk from state 2 to 4 for only three surfaces. The lack
of significance for the rest of the surfaces might be due to
the small numbers in the analyses.
The risk of developing new enamel lesions (from state 0
to 2) was significantly higher for most of the surfaces in in-
dividuals with DMFSappr c1 compared with those having
DMFSappr = 0–1 at the age of 11–12 years. This finding sup-
ports the results of Gröndahl et al. [1984], who reported that
13-year-olds with a DMFSappr c0 developed significantly
more new approximal dentin lesions during the following 3
years than those with a DMFSappr = 0 at 13 years. However,
the corresponding relative risk estimates of dentin caries
rates (states 2–4) did not reach statistical significance for
any of the surfaces.
How representative are the present data? Recently,
Marthaler et al. [1996] presented the average DMFT values
in 12-year-olds from a large number of countries and regions
in Europe from 1991 to 1995. The majority reported DMFT
values equal to or lower than the baseline DMFT at 12 (=
3.2) in the present cohort, suggesting that the present results
could be applicable to many contemporary populations.
It should be emphasized though that caries progression
in the outer part of the dentin could be assessed in this study
because it was undertaken in a population with a generally
low risk and low caries prevalence with access to organized
and regular dental care and that a special procedure was
used whereby a consensus at the clinic preceded any deci-
sion to restore an approximal surface [Mejàre et al., 1998].
In conclusion, generally, the enamel caries rate was low
(3.9 new enamel lesions/100 tooth surface-years) and pro-
gression of approximal enamel caries was a slow process.
Although progression was notably faster once the lesions
had reached the dentin, 50% of the lesions at the enamel-
dentin border did not show any radiographically visible
progression in the dentin after 3.1 years. Individuals with a
DMFSappr c1 at the age of 11–12 had a risk of developing
new enamel lesions that was 2.5 times greater than for those
with a DMFSappr = 0–1. A comparison of the age group of
11–16 with the age group of 17–22 years showed a signifi-
cantly faster caries rate from enamel to dentin (from state 2
to 4) for a few of the surfaces in the younger age group.
Acknowledgements
This study was supported by the Public Dental Health Service in
Stockholm, the Mayflower Foundation and the Swedish Patent Rev-
enue Research Fund.
200.89.68.56 - 3/31/2015 4:58:38 PM
Caries Res 1999;33:93–100 99
Universidad de Chile
Downloaded by:
 References
Bille J, Carstens JC: Approximal caries progres-
sion in 13–15-year-old Danish children. Acta
Odontol Scand 1989;47:347–354.
Carlos JP, Gittelsohn AM: Longitudinal studies of
the natural history of caries. II. A life-table
study of caries incidence in the permanent
teeth. Arch Oral Biol 1965;10:739–751.
Chestnut IG, Schäfer F, Jacobson APM, Stephen
KW: Incremental susceptibility of individual
tooth surfaces to dental caries in Scottish ado-
lescents. Community Dent Oral Epidemiol
1996;24:11–16.
Cohen J: A coefficient of agreement for nominal
scales. Educ Psychol Meas 1960;20:37–46.
Emslie RD: Radiographic assessment of approxi-
mal caries . J Dent Res 1959;38:1225–1226.
Gröndahl H-G, Andersson B, Torstensson T:
Caries increment and progression in teenagers
when using a prevention- rather than restora-
tion-oriented treatment strategy. Swed Dent J
1984;8:237–242.
Gröndahl H-G, Hollender L: Dental caries and
restorations. IV. A six-year longitudinal study
of the caries increment of proximal surfaces.
Swed Dent J 1979;3:47–55.
Gröndahl H-G, Hollender L, Malmcrona E,
Sundquist B: Dental caries and restorations in
teenagers. Swed Dent J 1977;1:45–50.
Hargreaves A: The clinical eruption of the perma-
nent teeth and observations noted during this
period. Odont Rev 1958;9:281–286.
Haugejorden O, Slack GL: Progression of approx-
imal caries in relation to radiographic scoring
codes: A pilot study. Acta Odontol Scand
1975;33:211–217.
Heidmann J, Hölund U, Poulsen S: Changing crite-
ria for restorative treatment of approximal
caries over a 10-year period. Caries Res
1987;21:460–463.
Holst D, Gjermo P, Rise J: Change in criteria for
treatment of approximal caries (abstract 36).
J Dent Res 1986;65:609.
100 Caries Res 1999;33:93–100
Hujoel PP, Isokangas PJ, Tiekso J, Davis S, Lam-
ont RJ, DeRouen TA, Mäkinen KK: A reanaly-
sis of caries rates in a preventive trial using
Poisson regression models. J Dent Res 1994;
73:573–579.
Kaplan EL, Meier P: Nonparametric estimation
from incomplete observations. J Am Stat As-
soc 1958;53:457–481.
Kay EJ, Nuttal NM, Knill-Jones R: Restorative
treatment thresholds and agreement in deci-
sion-making. Community Dent Oral Epidemi-
ol 1992;20:265–268.
Kleinbaum DG, Kupper LL, Morgenstern H: Epi-
demiologic Research. USA, Van Nostrand
Reinhold, 1982, chapt 6.
Kohlemainen L, Rytmömaa J: Increment of dental
caries among Finnish dental students during a
period of 2 years. Community Dental Oral Epi-
demiol 1977;5:140–144.
Kreiborg S, Rasmussen P, Thesleff I: Normal den-
tal and occlusal development; in Koch G,
Modéer T, Poulsen S, Rasmussen P (eds): Pe-
dodontics – A Clinical Approach. Copenhagen,
Munksgaard, 1994, p 50.
Lervik T, Haugejorden O, Aas C: Progression of
posterior approximal carious lesions in Norwe-
gian teenagers from 1982 to 1986. Acta Odon-
tol Scand 1990;48:223–227.
Lith A, Pettersson LG, Gröndahl H-G: Radio-
graphic study of approximal restorative treat-
ment in children and adolescents in two
Swedish communities differing in caries
prevalence. Community Dent Oral Epidemiol
1995;23:211–216.
Marthaler TM, O’Mullane DM, Vribic V: The
prevalence of dental caries in Europe
1990–1995: Symposium report. Caries Res
1996;30:237–255.
Mejàre I, Källestål C, Stenlund H, Johansson H:
Caries development from 11 to 22 years of age:
A prospective radiographic study. Prevalence
and distribution. Caries Res 1998;32:10–16.
Miettinen OS: Estimability and estimation in case-
referent studies. Am J Epidemiol 1976;103:
226–235.
Nuttall NM, Pitts NB: Restorative treatment
thresholds reported to be used by dentists in
Scotland. Br Dent J 1990;169:119–126.
Pine C, ten Bosch JJ: Dynamics of and diagnostic
methods for detecting small carious lesions.
Caries Res 1996;30:381–388.
Pitts NB: Monitoring of caries progression in per-
manent and primary posterior proximal enam-
el by bitewing radiographs. Community Dent
Oral Epidemiol 1983;11:228–235.
Pliskin JS, Shwartz M, Gröndahl H-G, Boffa J: Re-
liability of coding depth of approximal carious
lesions from non-dependent interpretation of
serial bitewing radiographs. Community Dent
Oral Epidemiol 1984;12:366–370.
Shwartz M, Gröndahl H-G, Pliskin J, Boffa J: A
longitudinal analysis from bite-wing radio-
graphs of the rate of progression of approximal
carious lesions through human dental enamel.
Arch Oral Biol 1984a;29:529–536.
Shwartz M, Pliskin JS, Gröndahl H-G, Boffa J:
Use of the Kaplan-Meier estimate to reduce bi-
ases in estimating the rate of caries progres-
sion. Community Dent Oral Epidemiol
1984b;12:103–108.
Shwartz M, Pliskin JS, Gröndahl H-G, Boffa J:
Study design to reduce biases in estimating the
percentage of carious lesions that do not
progress within a time period. Community
Dent Oral Epidemiol 1984c;12:109–113.
Strang D: The Radiographic Assessment of Poste-
rior Interproximal Dental Caries – A Longitu-
dinal Analysis, MSc thesis, Glasgow, 1990, ap-
pendix 1.
200.89.68.56 - 3/31/2015 4:58:38 PM
Mejàre/Källestål/Stenlund
Universidad de Chile
Downloaded by: