journal of dentistry 37 (2009) 712717

available at www.sciencedirect.com

journal homepage: www.intl.elsevierhealth.com/journals/jden

Restoration techniques and marginal overhang in Class II

composite resin restorations

B.A.C. Loomans *, N.J.M. Opdam, F.J.M. Roeters, E.M. Bronkhorst, M.C.D.N.J.M. Huysmans
College of Dental Science, Department of Preventive and Restorative Dentistry, Radboud University Nijmegen Medical Centre, The Netherlands

article info abstract

Article history: Objectives: The objective of the study was to compare in vitro interproximal overhang
Received 6 March 2009 formation of Class II composite resin restoration when using different matrix systems.
Received in revised form Methods: 240 lower left molar phantom head teeth with an MO-preparation were divided
19 May 2009 into 12 groups (n = 20). In six groups a circumferential matrix (Tofflemire X-thin matrix,
Accepted 20 May 2009 HaweNeos 1001-c, SuperCap) was used, combined with either a hand-instrument (PFI49 or
OptraContact) or separation ring (Composi-Tight Gold). In the other six groups two sectional
matrix systems were used (flexible and dead-soft), with three separation rings (Composi-
Keywords: Tight Gold, Contact Matrix, Palodent BiTine). Matrices were secured with wooden wedges
Interproximal overhang and preparations were restored with composite resin Clearfil AP-X (Kuraray) placed and
Class II polymerized in increments. After matrix removal overhang was measured on a standar-
Composite resin restoration dized digital macroscopic image in mm2. For analysis a multiple linear regression model was
used.
Results: Use of circumferential matrices resulted in less overhang than sectional matrices
( 0.85 mm2, p < 0.001). A flexible matrix led to less overhang than dead-soft matrices
( 0.54 mm2, p < 0.001), and no difference was found between straight and pre-contoured
matrices ( p = 0.945). The insertion of the OptraContact resulted in a much increased over-
hang of 2.54 mm2 ( p < 0.001). The Composi-Tight Gold and the Contact Matrix System rings
resulted in less overhang, 0.69 and 0.68 mm2, respectively (both p < 0.001), whereas the
Palodent BiTine ring did not.
Conclusions: Use of circumferential matrices or sectional flexible matrices resulted in the
least marginal overhang when combined with a Contact Matrix separation ring or a
Composi-Tight Gold ring.
# 2009 Elsevier Ltd. All rights reserved.

1. Introduction such iatrogenic factors and the pathogenesis of local period-

From literature it is known that overhanging restorations
promote gingivitis or lead to periodontal diseases due to local
accumulation of bacterial plaque rather than resulting in
mechanical irritation. Epidemiological and clinical experi-
mental studies have demonstrated close associations between
ontal lesions.17 However, all these studies have been
performed in a time where only amalgam was placed and
literature on the effect of composite resin overhang is scarce.
Formation of cervical overhang of the restoration is a
possible risk when placing Class II restorations. Overhang
formation is considered to be related to the type of restoration

* Corresponding author at: Radboud University Nijmegen Medical Centre, College of Dental Sciences, P.O. Box 9101, 6500 HB, Nijmegen, The
Netherlands. Tel.: +31 24 3616410; fax: +31 24 3540265.
E-mail addresses: b.loomans@dent.umcn.nl (B.A.C. Loomans), n.opdam@dent.umcn.nl (N.J.M. Opdam), j.roeters@dent.umcn.nl
(F.J.M. Roeters), e.bronkhorst@dent.umcn.nl (E.M. Bronkhorst), m.c.d.n.j.m.huysmans@dent.umcn.nl (M.C.D.N.J.M. Huysmans).
0300-5712/$ see front matter # 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jdent.2009.05.025
 journal of dentistry 37 (2009) 712717
material and matrix technique. Frankenberger et al. (1999)
showed that the use of flowable materials always led to higher
percentages of marginal overhangs in beveled cavities in vitro.
Moreover, higher viscous materials resulted in higher per-
centages of underfilled margins of beveled box-shaped
cavities.8 From a clinical study it was found that the majority
of approximal composite resin restorations presented
marginal overhang, which was also related to the anatomy
of the restored tooth.9 A study on the effect of different
matrix systems showed more overhang formation at the
margins when using plastic instead of metal matrices.10 In this
study the plastic matrices were combined with plastic
(reflective) wedges. These wedges are very stiff and lack the
ability of wooden wedges to adapt themselves to the natural
anatomic tooth contour. As a result, these wedges can make
contact to the matrix placed on the tooth at only one point
and permit the development of large gaps between the matrix
and the tooth at the critical cervical cavity margin and
generates substantial overhang formation during restoration
procedures. It is obvious that such an overhang at the
approximal part will be difficult to detect and even more
difficult to remove.
One of the goals in restorative dentistry is to re-establish a
tight proximal contact together with an optimal marginal
adaptation without overhang. In the past, good proximal
contacts were difficult to create with composite resin, as this
material cannot be condensed like dental amalgam. Now, with
the help of special separation rings, tight proximal contacts
can be established under in vitro and in vivo conditions.11,12
Besides the proximal contact tightness also the proximal
contour of the restoration is important. A well-contoured
proximal surface may help to prevent food impaction and will
facilitate interdental cleaning, both important factors to
maintain healthy interdental papillae.13 These anatomically
correct contoured restorations can be achieved when pre-
contoured (sectional or circumferential) matrix bands are used
for the restorative procedure.
The effect is of these modern matrix systems combined
with separation rings on the formation of interproximal
marginal overhang has not yet been evaluated. Therefore, the
objective of this study was to compare in vitro interproximal
overhang formation of Class II composite resin restoration
using different matrix systems.
2. Materials and methods
To simulate the clinical situation and to standardize the
restorative procedure a manikin model (KaVo, Dental,
Biberach, Germany) was used. For the experiments the contact
site between left second premolar and first molar in the lower
jaw was selected. In the molar tooth an MO-preparation was
made with a proximal box cavity of 5.0 mm in bucco-lingual,
6.0 mm in occlusalgingival and 1.3 mm in mesialdistal
dimension. The occlusal step was 4.5 mm in buccallingual
width, 2.5 mm deep and 6.0 mm in mesialdistal width. This
simulated an amalgam replacement situation. The margins of
the box were prepared without a bevel. Using a copy-milling
machine (Celay, Mikrona Technologie AG, Spreitenbach,
Switzerland) the model was replicated, resulting in 240
713
identical preparations in artificial first molars. The prepared
teeth were apically equipped with a stem-like anchoring
system, which allowed some mobility of the tooth simulating
normal physiological tooth mobility. Teeth were divided in 12
different groups (n = 20), each assigned to a specific matrix
type as shown in Table 1. Table 2 summarizes the product
profiles, LOT numbers and the characteristics of the materials
used in the study.
Matrix bands were secured interdentally from the buccal
side with wooden wedges (Slim-Jim, Wizard wedge) and in
groups 4 and 612 a separation ring was placed. Then in all
groups the contact area was burnished with a hand-instru-
ment (PFI49) so that no visual space was left between matrix
and adjacent tooth. All cavities were restored with an
adhesive and a hybrid composite (Clearfil Photo Bond and
Clearfil AP-X; Kuraray Medical, Tokyo, Japan). The adhesive
system was mixed and applied in the preparation, gently air-
dried and cured for 10 s with a halogen polymerization
unit (PolyLux II, KaVo, light intensity 600 mW/cm2). Subse-
quently, the composite resin was injected from the compule
into the cavity in two horizontal increments. In groups 13 a
hand-instrument (PFI49 or OptraContact) was used during
polymerization of the first layer to apply pressure towards
the distal contact area of the second premolar in order to
obtain a solid proximal contact. Each increment was
cured separately for 20 s from the occlusal surface. After
removal of the matrix, restorations were post-cured for 20 s
from the buccal and lingual side. Restorations were not
finished or adjusted in order to prevent changes of the
proximal surface. All restorations were placed by one
operator and all measurements were performed blind by
an independent observer.
Following the restorative procedure each tooth was
removed from the manikin model and placed horizontally
in a special mould made of polyvinylsiloxane (Express Putty
STD, 3 M ESPE). Using a stereomicroscope (Leica MZ 12)
standardized digital images were made of the proximal
surface with a magnification of 7.89, with the box placed
horizontally. Leica Qwin software was used to measure
digitally the total proximal restoration surface (mm2) by
marking the border of the restoration on the digital image.
Table 1 Experimental groups in the study.
Groups Matrix system
1 Tofflemire retainer + Tofflemire X-thin matrix + PFI49
2 Tofflemire retainer + 1101-c matrix + OptraContact
3 Tofflemire retainer + 1101-c matrix + PFI49
4 Tofflemire retainer + 1101-c matrix + Composi-Tight
Gold ring
5 SuperCap matrix
6 SuperCap matrix + Composi-Tight Gold ring
7 Contact Matrix System Stiff Flex matrix + Composi-Tight
Gold ring
8 Palodent Standard Matrix + Composi-Tight Gold ring
9 Contact Matrix System Stiff Flex matrix + Contact
Matrix ring
10 Palodent Standard Matrix + Contact Matrix ring
11 Contact Matrix System Stiff Flex matrix + Palodent
BiTine ring
12 Palodent Standard Matrix + Palodent BiTine ring
714 journal of dentistry 37 (2009) 712717

Table 2 Materials used in the study.

Materials Metal characteristic Matrix shape Manufacturers LOT

Standard Tofflemire Circumferential flexible Straight Produits Dentaire SA,

Matrix (Straight X-thin)
Hawe Contoured
Tofflemire-bands (1001-c)
Adapt SuperCap Matrix
(nr. 2162)
Contact Matrix (Stiff Flex)
Palodent Matrix system
(Standard Matrices)
Tofflemire retainer
Composi-Tight Gold
(AU 400)
Contact Matrix ring
(Outward rings)
Palodent BiTine ring
(Type I Round)
Wizard wedges
(Slim-Jim)
OptraContact Contact-point
instrument
PFI 49
Clearfil Photo Bond Catalyst
and Universal
Clearfil AP-X (PLT colour A3)
Vevey, Switzerland
Circumferential flexible Pre-contoured KerrHawe SA, Bioggio, 0090041050505
Switzerland
Circumferential flexible Pre-contoured KerrHawe SA, Bioggio, 0080041050505
Switzerland
Sectional flexible Pre-contoured Danville Materials, 89434
San Ramon, CA, USA
Sectional dead-soft Pre-contoured Dentsply Caulk, 559110
Milford, DE, USA
Produits Dentaire SA,
Vevey, Switzerland
Garrison Dental Solutions, 18884370032
Spring Lake, MI, USA
Danville Materials, 89507
San Ramon, CA, USA
Dentsply Caulk, Milford, 001204
DE, USA
Waterpik Technologies, 1672
CO, USA
Ivoclar Vivadent, Schaan, G14484
Liechenstein
Hu-Friedy, Chicago, IL, USA
Kuraray Medical, Osaka, Japan 41164
Kuraray Medical, Osaka, Japan 0068A

Only the marginal overhang in the cervical area of the box was
measured (50% box height: 3.0 mm  5.0 mm) (Fig. 1).
As the preparations were identical it was possible for the
software to automatically mark the cut-off point on the 50%
box height and only to include the area beneath this cut-off
line. Finally, the surface of the preparation (15.0 mm2) was
subtracted from the total recorded and marked surface,
resulting in the marginal overhang. The marginal overhang
for the buccal and the lingual part of the box was recorded
separately, in order to investigate the effect of wedge
direction. Data were statistically analyzed using SPSS 14.0.
To determine the effect of the experimental variables on the
marginal overhang, a multiple linear regression model was
constructed. In this model, the experimental conditions (type
of matrix and type of separation device) were included.
3. Results
Mean marginal overhang of the restorations is presented in
Fig. 2, together with standard deviations. The groups are
ordered from (top to bottom or left to right) in order of
increasing total overhang. Differences between groups were
large and total overhang ranged between 0.2 mm2 and
3.2 mm2 or between 2% and 21% of total restoration area.
Residuals of the multiple linear regression model were
normally distributed. The measurement error was established
by re-measuring five random specimens eight times in
average measurement error of 0.0014 mm2, varying between
0.0007 mm2 and 0.0024 mm2.
Whereas the circumferential matrices showed the lowest
overhang, in combination with the OptraContact instrument,

Fig. 1 Procedure of measuring the proximal marginal overhang using a stereomicroscope (7.89T); (a) only the marginal
overhang in the cervical area of the box was measured (50% box height), (b) marking the border of the restoration on the
digital image, (c) the surface of the preparation (15.0 mm2) was subtracted from the marked surface, resulting in the
marginal overhang.
 journal of dentistry 37 (2009) 712717 715

Fig. 2 Recorded proximal marginal overhang (total, buccal and lingual) of composite resin (mm2) together with the standard
deviation (S.D.).

the highest overhang was found. Table 3 presents the multiple
linear regression model. Buccal and lingual overhang showed
no significant differences, so total overhang was used as the
outcome variable. To predict the influence of other indepen-
dent variables, such as use of different separation rings or use
of different matrices on the obtained marginal overhang,
several combinations can be made by adding the impact
factors (Betas) of these variables. The adjusted R2 of the
regression model was 0.779.
The mean marginal overhang over all groups was regarded
as the reference group. Use of a circumferential matrix
resulted in less marginal overhang compared to the sectional
matrices ( 0.85 mm2, p < 0.001). Also the use of a flexible
matrix led to less overhang compared to the use of dead-soft
matrices ( 0.54 mm2, p < 0.001). No statistically significant
difference was found between the straight and pre-contoured
matrices ( p = 0.945). The insertion of the OptraContact device
during polymerization of the first layer of composite resulted
in the largest increase in marginal overhang of 2.54 mm2
( p < 0.001), whereas the use of the PFI49 resulted in a
decreased overhang ( p < 0.012).
The use of the Composi-Tight Gold and the Contact
Matrix System separation rings resulted in less overhang,
0.69 and 0.68 mm2, respectively (both p < 0.001), whereas
the Palodent BiTine ring had no effect on obtained overhang
( p = 0.532).
4. Discussion
In this study significant differences in the amount of
interproximal marginal overhang were found when using
different matrix systems. Differences in overhang were
related to the use of different separation rings, matrix bands
as well as the type of hand-instruments.
The methodology for the present study was partially of a
new design. For simulation of clinical conditions, the prepared
teeth were mounted in a phantom head and were not polished
after finishing in order to avoid interference with the
measurements. Overhang which is present at the occlusal
region of the restoration is relatively easy to remove or adjust,
however interproximal overhang at the cervical region will be
much more difficult or even impossible to remove. Therefore,
it was decided to include only the cervical area, 50% of total
box-height, of the restoration. The advantage of an in vitro
model is that it gives well-controlled experimental conditions

Table 3 The multiple linear regression model. In this model mean overhang over all groups was regarded as the
reference group. The adjusted R2 of the regression model was 0.779.
95% CI for Beta

Included Beta (mm2) p-Value Lower bound Upper bound

(Constant) 2.092 <0.001 1.713 2.470

Circumferential (1), Sectional (0) 0.848 <0.001 1.023 0.673
Flexible (1), Dead-Soft (0) 0.541 <0.001 0.670 0.412
Pre-contoured (1), Straight (0) 0.010 0.945 0.279 0.260
OptraContact 2.536 <0.001 2.230 2.841
PFI49 0.390 0.012 0.696 0.085
Composi-Tight Gold separation ring 0.687 <0.001 0.937 0.436
Contact Matrix System separation ring 0.682 <0.001 0.983 0.381
Palodent BiTine System separation ring 0.095 0.532 0.396 0.205
716 journal of dentistry 37 (2009) 712717
but it fails to account for the complexities of conditions that
are found under in vivo conditions. Teeth in the model had
some mobility simulating a normal physiological tooth
mobility; however, results might be somewhat influenced if
this mobility would be different. Therefore, the results found
in this study may under-estimate the interproximal overhang
presented in vivo.
The differences found between the matrix bands, flexible
or dead-soft, may be explained by the material characteristics
of these matrices. The flexible matrices demonstrate a more
elastic deformation during manipulation, whereas the dead-
soft matrices show a more plastic deformation. When the
tines of the separation ring come into contact with the matrix,
the dead-soft matrix will be more easily deformed and dented
than the more stiff, flexible, matrix, where the tines are more
likely to slide over the metal instead of squeezing it. Thus, the
flexible matrix is more likely to be pulled against the outline
than the dead-soft matrix.
A significant effect was found of the type of separation ring
used. The Contact Matrix ring and the Composi-Tight Gold
ring reduced overhang, whereas the Palodent BiTine ring did
not. This finding may be the result of the different ways in
which the tines of the rings touch the matrix and tooth. The
tines of the Palodent BiTine ring are parallel to each other and
relatively shorter than the other two rings (Composi-Tight
Gold and Contact Matrix). Therefore, the BiTine ring engages
the tooth more occlusally compared to the other two rings,
which might result in larger gaps between the preparation and
matrix at the cervical area.
When placing separation rings, several application techni-
ques are applicable to keep the ring in place. For example, in
case the clinical crown height is too low to place the ring
occlusally from the wedge, the ring can be placed next to the
wedge in order to give it more retention. In this study the rings
were all placed identically, as it is still unknown what the effect
is of different positioning methods on the formation of marginal
overhang. This aspect needs to be further investigated.
In this study a relatively low viscous syringable composite
(Clearfil AP-X) has been used and the effect of different
composite resin viscosities has not been investigated. In a
previous in vitro study it was shown that the use of lower viscous
composites (flowables) led to more marginal overhang com-
pared to higher viscous materials. On the other hand, higher
viscous materials (packables) led to more underfilled margins.8
The influence of wedge placement direction was investi-
gated by measuring the overhang for the buccal and lingual
side separately. As the wedge was placed from the buccal side,
it would be plausible that a better cervical adaptation would be
found at this side, resulting in less overhang formation.
However, from this study no statistically significant differ-
ences were found between the overhang at the cervico-buccal
and cervico-lingual side for any of the systems used.
Moreover, it was found that the circumferential matrices
resulted in less overhang compared to sectional matrices.
Circumferential matrices, placed in a Tofflemire retainer, were
firmly secured by the retainer resulting in a better adaptation
of the matrix to the tooth than the sectional matrices.
Unfortunately, these circumferential matrices, especially
when used without separation rings, hinder the reconstruc-
tion of a tight proximal contact.11 The use of the hand-
instrument OptraContact resulted in dramatically increased
overhang compared to the use of the hand-instrument PFI49,
which reduced overhang. This result may be explained by the
different shape and use of the instrument. The PFI49 has the
form of a drop of water, is round and relatively small whereas
the OptraContact is fork-like has two small tines and is wider
than the PFI49. When inserting the hand-instrument in the
uncured composite the OptraContact will push the matrix
over a broader area in the direction of the adjacent tooth,
pulling the matrix away from the margin of the preparation,
with a gapand therefore marginal overhangas a result.
It was noted in this study that all of the examined
dental restorations displayed material overhang and most of
these overhangs were located in the cervical area of the
restoration.
Despite all efforts to prevent interproximal overhang, it is
clear from literature9,10 as well as this study, that complete
prevention of interproximal overhang is almost impossible.
From all studies dealing with marginal overhang most were
related to amalgam restorations and also no distinction was
made between a smooth or a rough and uneven overhang.110
Therefore, results found in these studies may not be
generally valid for composite resin restorations. Depending on
the clinical appearance of the composite resin overhang, there
might be a different approach for the possible treatment of it.
When the overhang is smooth and continuous (a flash) with
the outline of the tooth, it may not promote plaque or calculus
accumulation, and it can be doubted if such overhang is a
clinical problem that has to be removed. However, when this
flash dislodge it might result in a rough margin and possible
irritate the epithelial attachment. Such a rough and bulky
overhang might form a retention place for plaque accumula-
tion promote (secondary) caries and/or periodontal disease.
Unfortunately, up till now the exact clinical relevance and
effect of such an interproximal overhang (smooth or rough) on
periodontal complications or the occurrence of secondary
caries is still unknown and has to be subject of further
investigation.
5. Conclusions
Within the limitations of the in vitro study, it can be concluded
that:
All matrix systems resulted in marginal overhang of Class II
composite resin restorations.
Circumferential matrices resulted in less overhang com-
pared to sectional matrices.
Flexible sectional matrix systems provided less overhang
compared to dead-soft matrices.
The use of a Contact Matrix separation ring or a Composi-
Tight Gold ring resulted in less overhang, whereas Palodent
BiTine ring did not.
Acknowledgements
The authors would like to thank Ms. Jian Phialy and Ms. Juliet
van Bochove for their contribution to this research.
 journal of dentistry 37 (2009) 712717
references
1. Gilmore N, Sheiham A. Overhanging dental restorations and
periodontal disease. Journal of Periodontology 1971;42:812.
2. Tervonen T, Ainamo J. Relative influence of calculus and
overhangs of fillings on the frequency of score 2 of the
CPITN. Community Dentistry and Oral Epidemiology
1986;14:136217.
3. Jansson L, Ehnevid H, Lindskog S, Blomlof L. Proximal
restorations and periodontal status. Journal of Clinical
Periodontology 1994;21:57782.
4. Hakkarainen K, Ainamo J. Influence of overhanging
posterior tooth restorations on alveolar bone height
in adults. Journal of Clinical Periodontology 1980;7:11420.
5. Pack ARC, Coxhead LJ, McDonald BW. The prevalence of
overhanging margins in posterior amalgam restorations and
periodontal consequences. Journal of Clinical Periodontology
1990;17:14552.
6. Rodriguez-Ferrer HJ, Strahan JD, Newman HN. Effect of
gingival health or removing overhanging margins of
interproximal subgingival amalgam restorations. Journal of
Clinical Periodontology 1980;7:45762.
7. Than A, Duguid R, McKendrick AJW. Relationship between
restorations and the level or the periodontal attachment.
Journal of Clinical Periodontology 1982;9:193202.
717
8. Frankenberger R, Kramer N, Pelka M, Petschelt A. Internal
adaptation and overhang formation of direct Class II
composite restorations. Clinical Oral Investigation
1999;3:20815.
9. Opdam NJM, Roeters FJM, Feilzer AJ, Smale I.
A radiographic and scanning electron microscopic
study of approximal margins of Class II resin composite
restorations placed in vivo. Journal of Dentistry
1998;26:31927.
10. Mullejans R, Badawi MOF, Raab WHM, Lang H.
An in vitro comparison of metal and transparent
matrices used for bonded class II resin composite
restorations. Operative Dentistry 2003;28:
1226.
11. Loomans BAC, Opdam NJM, Roeters FJM, Bronkhorst EM,
Burgersdijk RCW, Dorfer CE. A randomized clinical trial on
proximal contacts of posterior composites. Journal of
Dentistry 2006;34:2927.
12. Loomans BAC, Opdam NJM, Roeters FJM, Bronkhorst EM,
Plasschaert AJM. Influence of composite resin consistency
and placement techniques on proximal contact tightness of
Class II restorations. Journal of Adhesive Dentistry 2006;8:
30510.
13. Barnes DM, Blank LW, Thompson VP, Holston AM,
Gingel JC. A 5- and 8-year clinical evaluation of a
posterior composite resin. Quintessence International
1991;22:14351.