Journal of Oral Rehabilitation
Journal of Oral Rehabilitation 2011 38; 469–474
Characteristics of non-carious cervical lesions – an ex vivo
study using micro computed tomography
B. HUR*, H-C. KIM*, J-K. PARK* & A. VERSLUIS†
of Dentistry, Pusan National University, Yangsan, Korea and †Department of Bioscience Research, College of Dentistry, University of Tennessee
Health Science Center, Memphis, TN, USA
SUMMARY The aetiology of non-carious cervical
lesions (NCCLs) is not well understood and still
controversial. The aim of this ex-vivo study was to
examine the morphological characteristics of NCCLs
for clinical evidence of enamel loss above the
cemento-enamel junction (CEJ) as suggested by the
abfraction theory. Fifty extracted human teeth with
various types of NCCLs were collected and scanned
by micro computed tomography. The reconstructed
three-dimensional models were evaluated from
multiple aspects, including longitudinal cross-sec-
tion series. The location of internal line angle and
proximal exits of the lesions were evaluated in
relation to the level of CEJ. The coronal margins of
the lesions were inspected for evidence of enamel
Introduction
Non-carious cervical lesions (NCCLs), characterised by
the loss of hard tissue at the cemento-enamel junction
(CEJ) in the absence of caries, are a condition com-
monly encountered in dental practice (1). It is generally
accepted that initiation and progression of NCCLs have
a multi-factorial aetiology, but the relative contribu-
tions of the various processes still remain unresolved
and a source of scientific debates. Currently, the most
widely accepted causes of NCCLs are abfraction and
abrasion, while other mechanisms such as erosion and
corrosion have also been proposed (2).
The abfraction hypothesis is based on a biomechan-
ical theory in which tensile stress concentrations at the
cervical regions of teeth, caused by tooth flexure during
heavy occlusal loading, lead to micro-crack formation
ª 2010 Blackwell Publishing Ltd
*Department of Conservative Dentistry, School
loss above the CEJ using the bucco-lingual longitu-
dinal sectional images. Coronal margins of the
lesions were located along and ⁄ or under the CEJ
for all of the 50 samples. In most of the lesions,
regardless of lesion type, the proximal exits and
internal line angles were located below the CEJ. This
study did not detect clinical evidence of enamel loss
above the occlusal margin of NCCLs as would have
been expected according to the general abfraction
mechanism.
KEYWORDS: non-carious cervical lesions, abfraction,
abrasion, cemento-enamel junction, micro-CT
Accepted for publication 19 September 2010
cervically by disrupting bonds between hydroxyapatite
crystals in enamel and dentine (3). Some studies (4–6)
support the abfraction theory by demonstrating
susceptibility of enamel fracture (loss) above the CEJ
through finite element analysis (FEA). Consistent
with the abfraction theory, coronal margins of NCCLs
are therefore often assumed to be located above the
CEJ (7, 8).
If abfraction is the predominant mechanism, enamel
loss above the CEJ should be prevalent clinically.
Clinical manifestations of NCCLs have been studied by
clinical examination (9) and by optical or scanning
electron microscopic examination of replicas (10) or
extracted teeth (11). Pintado et al. (12) followed the
development of tooth structure loss of NCCLs in one
patient over a 14-year period. However, although
general lesion morphologies have thus been identified
doi: 10.1111/j.1365-2842.2010.02172.x
470 B . H U R et al.
(a) (b) (c)
(d) (e)
or even quantified in these studies, they did not
describe the tooth structure loss for enamel specifically
(13, 14).
The aim of this study was to examine the morpho-
logical characteristics of NCCLs for clinical evidence of
enamel loss as suggested by current abfraction theory at
and ⁄ or above the CEJ. Micro Computed Tomography
(micro-CT) imaging of extracted teeth with NCCLs will
be used. Micro-CT allows the complete structure of an
object to be stored and examined to give all internal
dimensions and the precise size, shape and location of
any internal feature or defect in a three-dimensional
(3D) reconstruction. With this method, lesion charac-
teristics can be inspected by imaging at various view
angles and cross sections.
Material and methods
Fifty extracted human teeth with various characteristics
of NCCLs were collected. Samples with intact NCCLs
were selected randomly, regardless of tooth type, lesion
shape or lesion size. The teeth were stored in 0Æ5%
sodium hypochlorite solution for dissolution of soft
tissues and decontamination. The selected samples
comprised of three molars, 33 premolars and 14
anteriors. The samples were scanned by micro-CT* at
30 micro-metre intervals. The scanned images were
reconstructed three-dimensionally using an image
*HMXCT 225; X-Tek system Ltd, Tring, Herts, UK.
Fig. 1. Representative sample of
premolar having non-carious cervi-
cal lesion and morphological charac-
terizations. (a–c) Three-dimensional
images (a, mesial aspect; b, buccal
aspect; c, distal aspect) and (d)
bucco-lingual longitudinal sectional
image of the lesion. (e) Illustrates the
definition of ‘blunt enamel’ at the
coronal margin.
reconstruction program†. The reconstructed model of
each tooth (700 · 700 · 700 voxels) was evaluated
from multiple 3D aspects, including longitudinal
cross-sectional slice images. The NCCLs’ morphological
characteristics, including enamel loss, were evaluated
using the 3D mid-buccal, mesio-buccal and disto-buccal
aspects as well as the longitudinal X-ray images
captured using the 3D reconstruction program (Fig. 1).
The location of internal line angle and proximal exits
of the lesions were evaluated in relation to the level of
the CEJ. The coronal margins of the lesions were
inspected for evidence of enamel loss (blunt enamel
margin) above the CEJ using the bucco-lingual longi-
tudinal sectional images at the most prominent (mid-)
buccal surfaces (Fig. 1). The shapes of the lesions were
characterised by the presence of sharp internal line
angles at the deepest portion of the lesion. The lesions
with sharp internal line angles were classified as wedge
shaped, while the lesions with broad and round line
angles were classified as saucer shapes. Lesion sizes,
representing the degree of lesion progression, were
quantified by measuring their height and depth using
the longitudinal sectional image.
Results
Three types of lesions were found: wedge shape, saucer
shape and mixed shape. Figure 2 shows mid-buccal
†
Ez3D2009; E-WOO Technology Co., Seoul, Korea.
ª 2010 Blackwell Publishing Ltd
 NON-CARIOUS CERVICAL LESIONS – ABRASION OR ABFRACTION? 471

(a)

(b)

Fig. 2. Representative samples of

three types of the lesions: (a) wedge-
shaped lesion, (b) saucer-shaped
lesion and (c) mixed shape lesion.
Left column shows mid-buccal (c)
aspect, mid column shows proximal
aspect and the right column shows
buccolingual longitudinal sectional
images. Sectional images show the
intact sharp enamel margins (SEM)
at coronal margins of the lesions.

aspects, proximal aspects and buccolingual longitudinal margins at the CEJ, indicating intact CEJs. In a few
sections for a representative sample of each type. cases (6 ⁄ 50), the proximal exits were located near the
Wedge-shaped lesions (20 ⁄ 50) had a sharp internal CEJ (Fig. 3) and in these cases they still had their
line angle with a V-shape, while saucer shapes (21 ⁄ 50) internal line angles under the CEJ and an intact sharp
had a rounded internal line angle of C- or U-shape. enamel margin. Even in rare cases (2 of 20 wedge shape
Mixed shape lesions (9 ⁄ 50) showed flat cervical walls lesions) of double-wedge shapes which had severe
and semicircular occlusal walls. The mean height and tooth material loss (Fig. 3), a sharp enamel margin was
depth of all specimens are presented in Table 1. found along the CEJ.
Regardless of tooth type, lesion shape or lesion size,
none of the coronal margins of the lesions were located
Discussion
above the CEJ. The 3D micro-CT models showed
coronal margins of lesion cavities near the CEJ at The formation of NCCLs has long intrigued researchers
buccal area and below ⁄ near the CEJ at the proximal and clinicians. Various hypotheses have been proposed
area. In most of the lesions, regardless of lesion types, for its aetiology, but except for its multi-factorial
the proximal exits and internal line angles were located nature, no consensus exists about its mechanism. One
below the CEJ (Fig. 2). All coronal margins in the of the theories put forward is based on a biomechanical
longitudinal sectional images showed sharp enamel concept in which the cervical area of a tooth becomes a
fulcrum during occlusal function, bruxing and para-
Table 1. Sample composition and the cavity size (height and
depth; mm) of each type of non-carious cervical lesion
functional activity, causing tensile stresses in the area
where NCCLs occur. These stresses are thought to
Wedge Saucer Mixed disrupt the crystalline structure of the locally thin
Tooth type Dimension shape shape shape enamel and underlying dentin by cyclic fatigue, leading
Anteriors, Height 4Æ75 n = 3 2Æ82 n = 10 3Æ2 n = 1
to cracks. Ultimately, the enamel breaks away at the
n = 14 Depth 1Æ94 0Æ25 1Æ17 cervical margin and progressively exposes the dentin,
Premolars, Height 3Æ34 n = 17 2Æ74 n = 8 3Æ3 n = 8 where the process continues (3). This study examined if
n = 33 Depth 1Æ36 0Æ99 1Æ85 clinically occurring NCCLs offer evidence in their
Molars, Height – n=0 4Æ62 n = 3 – n=0 morphologies for the theorised description of the
n=3 Depth – 1Æ29 –
abfraction mechanism.

ª 2010 Blackwell Publishing Ltd

472 B . H U R et al.
(a) (b)
Morphological features of NCCLs have been
described as two distinct patterns, one with sharp
internal line angle, the other more rounded, by visually
inspecting the vertical bucco-lingual cross section of
clinical samples (15, 16). Being that NCCLs are saucer
or wedge shaped, they have been categorised by the
angle (sharp internal line angle or rounded smooth
angle) formed by the occlusal and gingival walls.
Levitch et al. (17) described the common types of
abrasive cervical lesions as simple flat-floored grooves,
defects that are C-shaped in cross section with rounded
floors, undercut defects with a flat cervical wall and a
semicircular occlusal wall, as well as typical V-shaped
grooves with oblique walls that intersect axially. All the
three distinct types of lesions categorised according to
these descriptions were identified in the samples
examined in our study, which thus presents a credible
variation of the three shapes and degrees of lesion
progression (Table 1).
None of the samples in the current study showed a
blunt enamel margin at the occlusal wall as would have
been expected according to the abfraction theory.
Initiation and propagation of enamel cracks in the
cervical area on which the abfraction concept is based
depend strongly on the presence of stresses induced by
occlusal loads. There is some clinical evidence (12, 18)
that supports the association between occlusal loads
and cervical wear, but the suggestion of enamel fracture
at the CEJ has been mainly based on FEA (19–21) and
laboratory studies (22), with little data to confirm such
enamel fracture clinically.
Finite element analysis is a versatile tool to study
stress conditions within complex structures, but these
Fig. 3. The proximal exits of
the lesion are located near the
cemento-enamel junction (CEJ).
(a) The longitudinal bucco-lingual
sections show that the intact and
sharp enamel margin (SEM) and the
internal line angle are located below
the CEJ. (b) A rare case of severe
tooth material loss in a wedge shape
non-carious cervical lesions. The
longitudinal bucco-lingual section
also shows an intact sharp enamel
margin along the CEJ.
conditions can become predetermined by presuming
NCCL shapes and locations, and by assumptions about
the boundary conditions (tooth loading and fixture).
Furthermore, fracture modelling of tooth tissues adds
an additional level of complexity. Clinical data such as
the current study are thus extremely valuable to
validate and guide FEA. Clinical data are equally
valuable for the validation of in vitro studies. Experi-
mental tooth-brushing simulation results have shown
an abrasive lesion that began apical to the CEJ and then
continued abrading dentin, undermining enamel and
eventually wearing away enamel (23). This experi-
mentally induced toothbrush abrasion was reported to
duplicate the classical clinical shapes such as wedge,
wedge ⁄ round and rounded, but in all cases, the original
CEJ was lost. The NCCLs examined in our ex vivo
samples generally appeared to be limited coronally by
the CEJ and extended from the CEJ to the root surfaces.
A recent clinical study by Nguyen et al. (14) made the
same observation. Although the cavity shapes produced
in the in vitro study thus appeared similar to clinical
lesions, they did not replicate the preserved enamel
above the CEJ shown in our study of clinically formed
lesions.
Based on our detailed ex vivo examination of clinical
NCCLs, we did not find evidence to support the loss of
cervical enamel as predicted by the theorised abfraction
mechanism. Recently, Daley et al. (10) also reported
that they could not find histopathological evidence of
abfraction. Bartlett and Shah (24) published a critical
review about NCCLs in which they concluded that
there was, as yet, insufficient evidence to support the
abfraction mechanism. Their literature review strongly
ª 2010 Blackwell Publishing Ltd
 NON-CARIOUS CERVICAL LESIONS – ABRASION OR ABFRACTION?
supported erosion and abrasion as dominant factors in
the development of wedge-shaped lesions along the
cervical margins of teeth. The location of a lesion on
exposed root surface has frequently been associated
with gingival recession or loss of attachment, either
concurrent with or prior to initiation of the abfraction
lesion (25). The ability to vary root exposure, tooth
juxtaposition, gingival contour, toothbrush firmness
and toothpaste abrasivity, along with exposing the CEJ
to the filaments during brushing, may all support an
abrasion theory for NCCL formation. The various
shapes of created NCCLs have been proposed to be
caused by toothbrush filament deflection (26). NCCL
shapes could also be related to the wear rates of
cementum and dentin, which are much higher than
enamel (27), which is about 19Æ9 times more resistant
to abrasion than dentin (28). These considerations may
hold a more apt explanation for the location of the
NCCLs observed in our study and the absence of
enamel loss than the abfraction theory. However,
long-term clinical studies based on a large number of
samples with NCCLs will be required to gather the
evidences to support the relationship of occlusal stress,
tooth brushing and gingival recession in the formation
of NCCLs.
Conclusions
Based on the observations of 50 ex vivo teeth with
clinically formed NCCLs, it can be concluded that
1 Regardless of tooth type, lesion shape or lesion size,
all lesions were located below the level of CEJ.
2 No clinical evidence of enamel fractures at coronal
margin of NCCLs was detected.
Acknowledgments
The authors thank Ms. Maria R. Pintado, BS, MPH,
Associate Professor, Minnesota Dental Research Center
for Biomaterials and Biomechanics, School of Dentistry,
University of Minnesota for her advice, encouragement
and writing assistance. Without her knowledge and
assistance, this study would not have been successful.
References
1. Lyttle HA, Sidhu N, Smyth B. A study of the classification and
treatment of noncarious cervical lesions by general practitio-
ners. J Prosthet Dent. 1998;79:342–346.
ª 2010 Blackwell Publishing Ltd
473
2. Michael JA, Townsend GC, Greenwood LF, Kaidonis JA.
Abfraction: separating fact from fiction. Aust Dent J.
2009;54:2–8.
3. Lee WC, Eakle WS. Possible role of tensile stress in the
etiology of cervical erosive lesions of teeth. J Prosthet Dent.
1984;52:374–380.
4. Lee HE, Lin CL, Wang CH, Cheng CH, Chang CH. Stresses at
the cervical lesion of maxillary premolar – a finite element
investigation. J Dent. 2002;30:283–290.
5. Dejak B, Młotkowski A, Romanowicz M. Finite element
analysis of stresses in molars during clenching and mastica-
tion. J Prosthet Dent. 2003;90:591–597.
6. Borcic J, Anic I, Smojver I, Catic A, Miletic I, Ribaric SP. 3D
finite element model and cervical lesion formation in normal
occlusion and in malocclusion. J Oral Rehabil. 2005;32:
504–510.
7. Yettram AL, Wright KW, Pickard HM. Finite element stress
analysis of the crowns of normal and restored teeth. J Dent
Res. 1976;55:1004–1011.
8. Palamara D, Palamara JE, Tyas MJ, Messer HH. Strain patterns
in cervical enamel of teeth subjected to occlusal loading. Dent
Mater. 2000;16:412–419.
9. Piotrowski BT, Gillette WB, Hancock EB. Examining the
prevalence and characteristics of abfraction like cervical
lesions in a population of U.S. veterans. J Am Dent Assoc.
2001;132:1694–1701.
10. Daley TJ, Harbrow DJ, Kahler B, Young WG. The cervical
wedge-shaped lesion in teeth: a light and electron microscopic
study. Aust Dent J. 2009;54:212–219.
11. Grippo JO. Abfractions: a new classification of hard tissue
lesions of teeth. J Esthet Dent. 1991;3:14–19.
12. Pintado MR, DeLong R, Ko C-C, Sakaguchi RL, Douglas WH.
Correlation of noncarious cervical lesion size and occlusal
wear in a single adult over a 14-year time span. J Prosthet
Dent. 2000;84:436–443.
13. Bernhardt O, Gesch D, Schwahn C, Mack F, Meyer G, John U
et al. Epidemiological evaluation of the multifactorial aetiol-
ogy of abfractions. J Oral Rehabil. 2006;33:17–25.
14. Nguyen C, Ranjitkar S, Kaidonis JA, Townsend GC. A
qualitative assessment of non-carious cervical lesions in
extracted human teeth. Aust Dent J. 2008;53:46–51.
15. Brady JM, Woody RD. Scanning microscopy of cervical
erosion. J Am Dent Assoc. 1977;94:726–729.
16. Aw TC, Lepe X, Johnson GH, Mancl L. Characteristics of
noncarious cervical lesions: a clinical investigation. J Am Dent
Assoc. 2002;133:725–733.
17. Levitch LC, Bader JD, Shugars DA, Heymann HO. Non-
carious cervical lesions. J Dent. 1994;22:195–207.
18. Madani AO, Ahmadian-Yazdi A. An investigation into the
relationship between noncarious cervical lesions and pre-
mature contacts. Cranio. 2005;23:10–15.
19. Rees JS. The effect of variation in occlusal loading on the
development of abfraction lesions: a finite element study.
J Oral Rehabil. 2002;29:188–193.
20. Tanaka M, Naito T, Yokota M, Kohno M. Finite element
analysis of the possible mechanism of cervical lesion forma-
tion by occlusal force. J Oral Rehabil. 2003;30:60–67.
474 B . H U R et al.
21. Dejak B, Mlotkowski A, Romanowicz M. Finite element
analysis of mechanism of cervical lesion formation in simu-
lated molars during mastication and parafunction. J Prosthet
Dent. 2005;94:520–529.
22. Whitehead SA, Wilson NH, Watts DC. Development of
noncarious cervical notch lesions in vitro. J Esthet Dent.
1999;11:332–337.
23. Litonjua LA, Andreana S, Bush PJ, Tobias TS, Cohen RE.
Wedged cervical lesions produced by toothbrushing. Am
J Dent. 2004;17:237–240.
24. Bartlett DW, Shah P. A critical review of non-carious cervical
(wear) lesions and the role of abfraction, erosion, and
abrasion. J Dent Res. 2006;85:306–312.
25. Lyons K. Aetiology of abfraction lesions. N Z Dent J.
2001;97:93–98.
26. Dzakovich JJ, Oslak RR. In vitro reproduction of noncarious
cervical lesions. J Prosthet Dent. 2008;100:1–10.
27. Litonjua LA, Andreana S, Cohen RE. Toothbrush abra-
sions and noncarious cervical lesions: evolving concepts.
Compend Contin Educ Dent. 2005;26:767–768, 770–774,
776.
28. Bishop K, Kelleher M, Briggs P, Joshi R. Wear now? An
update on the etiology of tooth wear. Quintessence Int.
1997;28:305–313.
Correspondence: Dr Hyeon-Cheol Kim, Department of Conservative
Dentistry, School of Dentistry, Pusan National University, Beomeo-Ri,
Mulgeum, Yangsan, Gyeongnam 626-810, Korea.
E-mail: golddent@pusan.ac.kr
ª 2010 Blackwell Publishing Ltd