ORIGINAL ARTICLE

The influence of PDL principal fibers in a

3-dimensional analysis of orthodontic
tooth movement
Haihong Qian, MSME,a Jie Chen, PhD,b and Thomas R. Katona, PhD, DMDb
West Lafayette and Indianapolis, Ind

The effects of mechanical loads on the tooth-alveolus complex are of particular concern in orthodontics. The
concepts of center of resistance (CRes) and center of rotation (CRot) are used to characterize tooth
responses to orthodontic loads. The mechanical environment (stresses and strains) associated with
orthodontic tooth movement is a unique model in bone adaptation physiology. Numerous finite element
models of varying complexity have been developed to calculate tooth movements and stress distributions
within the alveolar bone and the periodontal ligament (PDL). In general, the PDL has been idealized as a
homogeneous isotropic material. For this project, a 3-dimensional tooth/PDL/mandible/finite element model
was developed in which, for the first time in such an analysis, the PDL’s principal-fiber structure was also
incorporated. Parametric analyses showed that the fiber orientation and the mechanical properties do not
exert much influence on the locations of the CRes and the CRot and on the stress patterns within the bone
and the PDL matrix. However, the absence of principal fibers produces not only different stress magnitudes,
but also differences in stress patterns. Furthermore, the no-fiber–associated CRes and CRot are considerably
separated from the cluster of fiber-influenced centers. It was concluded that it may be more realistic to
incorporate “generic” principal fibers into finite element models than not to include them at all, despite the lack
of reliable information about fibers. (Am J Orthod Dentofacial Orthop 2001;120:272-9)

O
rthodontic tooth movement is often described
in terms of the location of the center of rotation
(CRot). The CRot is an imaginary point, not
necessarily on the tooth, about which the tooth can be
considered to have purely rotated.1 The center of resis-
tance (CRes) is defined as any point along the line of
action of a force that causes pure translation of the
tooth. If the line of action of a force passes through the
CRes, then the tooth will undergo pure translation—
that is, the CRot approaches infinity. In general, if the
line of action of a force does not go through the CRes,
then the tooth undergoes a combination of translation
and rotation. (Any combination of displacements can
be reduced to an equivalent pure rotation about some
CRot.)
From the Biomechanics and Biomaterials Research Center, Indiana University-
Purdue University, Indianapolis, Ind.
Supported by NIH/NIDR (DE11058).
aGraduate student in mechanical engineering, Purdue University, West
Lafayette, Ind.
bAssociate Professor of Orthodontics, Indiana University School of Dentistry,
and of Mechanical Engineering, Purdue University School of Engineering and
Technology.
Reprint requests to: Dr Thomas R. Katona, Department of Oral Facial Devel-
opment, Indiana University School of Dentistry, 1121 W Michigan St, Indi-
anapolis, IN 46202; e-mail, tkatona@iupui.edu.
Submitted, September 2000; revised and accepted, February 2001.
Copyright © 2001 by the American Association of Orthodontists.
0889-5406/2001/$35.00 + 0 8/1/116085
doi:10.1067/mod.2001.116085
272
In the bone research community, mechanical-load
transduction into bone response has been the subject of
intense investigation. Orthopedic long-bone models
predominate, although orthodontic models offer unique
and powerful alternatives. The periodontal ligament
(PDL) transmits functional and orthodontic forces to
the alveolar bone. It is believed that the resulting
stresses or strains within the PDL and the bone control
the bone-modeling process. Calculations using finite
element method (FEM) models of varying complexity
and verisimilitude have been used to investigate the
relationship between the changed mechanical environ-
ment and the bone response.2-6
Virtually all models have shared the assumption
that the PDL is homogenous and isotropic. But histo-
logic studies have clearly demonstrated that the PDL is
a fiber-reinforced structure.7-9 Presumably, the princi-
pal fibers resist tensile forces, whereas the other com-
ponents resist compressive stresses.
The purposes of this study were to investigate the
possible structural roles of the PDL’s principal fibers
in influencing the locations of the CRot and the CRes,
and the effects of the fibers on the stress fields within
the PDL and the adjacent bone. A 3-dimensional FE
model of the root/PDL/mandible structure was built,
and, for the first time in this type of study, the PDL’s
principal-fiber structure was also simulated with spe-
American Journal of Orthodontics and Dentofacial Orthopedics Qian, Chen, and Katona 273
Volume 120, Number 3

Fig 2. Finite element mesh of orthodontic model with

tooth, PDL, and mandible. Mandible consists of cortical
Fig 1. Schematic of mandible cross section. and cancellous bone.

Table I.Fiber density and area (estimates based on pic- Table III. Combinations of fiber properties used in
tures in Auyeung7) parametric calculations; in addition to 5 cases in Table,
Region Density (/100 µm) Area (µm2) results for traditional no-fiber case and another model
with only principal fibers (no-NPFM case; 45° and
Coronal and middle 22.1 9.0 E = 30 MPa) were obtained
Apical 13.2 16.6
Angulation (in degrees)

Young’s modulus (MPa) 30 45 60

Table II. Material properties
10 X
Young’s modulus
30 X X X
Material (MPa) Poisson’s ratio
90 X
Tooth 18,0005 0.30
Cortical bone of mandible 13,00015 0.30
Cancellous bone of mandible 1,00015 0.30
MATERIAL AND METHODS
PDL non-fibrous matrix (NPFM) 2 0.45
PDL fiber 10/30/90 0.35 Geometry and assumptions
The model is a simplified canine dog tooth/PDL/
mandible structure in which the tooth is regarded as
cial reinforced elements. (Some aspects of the current composed entirely of dentin. Cancellous bone forms the
project were attempted in an early FEM study,10 and interior of the mandibular body, and cortical bone cov-
the same approach has been used to model tooth erup- ers the mandible and forms the tooth socket (lamina
tion.11) A mesially directed orthodontic force was dura). The geometric approximations (Fig 1) were
applied to the crown of the tooth. Because of the obtained from another study in which the distal root of
paucity of information about fibers, a large-range para- a mandibular second premolar and its supporting bone
metric study of fiber orientation and fiber Young’s were serially sectioned and measured.
modulus of elasticity was performed. That is, with var- The PDL in the model contains 2 components: the
ious combinations of these fiber properties, the CRot principal (oblique and apical) fibers and everything else
and the CRes, and stresses within the PDL and the in the PDL space. The second component is called the
socket bone (lamina dura), were calculated as the “nonprincipal-fiber matrix” (NPFM). Including princi-
PDL’s principal fiber orientation and Young’s modulus pal fibers makes this model unique because previous
were parametrically altered. models generally included only the NPFM or the fibers,
274 Qian, Chen, and Katona
Fig 3. Cross section of finite element model. One of 2
forces applied to crown on lingual side is partly hidden.
(Because stresses of interest in PDL and adjacent bone
are relatively far removed from points of force application
on crown, anatomic detail of crown is superfluous.)
but not both. We followed traditional practice when we
assumed that the NPFM would behave as a homoge-
nous and isotropic material.
Although actual oblique fiber orientation and stiff-
ness vary along the root,8 the orientation and stiffness
were assumed to be constant, except at the apical
region. (The effects of this assumption are examined in
the parametric analysis.) Although the principal fibers
behave nonlinearly,12 we assumed that they are linearly
elastic. The absence of principal fiber stiffness in com-
pression is reflected in this model. Two FEM parame-
ters (fiber density and fiber intersection area) had to be
defined to numerically simulate the principal fibers.
These approximations (based on pictures from Auyeung
et al7) are shown in Table I.
The mechanical properties (Table II) were collated
from the literature and our previous work. Unfortu-
nately, there are no universally accepted bone property
values; the literature is replete with many possibilities.
However, for this study, a parametric analysis of the
effects of fiber variability, we believed that bone-
property changes had a minimal effect. This can be
rationalized in 2 ways: (1) bone stiffness values are
American Journal of Orthodontics and Dentofacial Orthopedics
September 2001
Fig 4. Principal fiber arrangement in apical PDL ele-
ments. In “single”-type element, 1 fiber in each element
links PDL-root and PDL-bone interfaces.
Fig 5. Oblique principal PDL fiber attachment away from
apical region. Reinforcements in “skew”-type elements
are arranged in layers whose angulations are defined by
f/L ratio.
orders of magnitude greater than the fiber values (Table
II), so that even relatively large changes in bone stiff-
ness are unlikely to have a significant effect; and (2)
the results are primarily analyzed in a relative sense. As
an example, an alteration in bone properties may
slightly move the calculated CRes locations, but the
relative differences in the locations computed for dif-
ferent fiber properties/orientations would be much less
affected.
Table III presents the combinations of fiber angula-
tions and stiffnesses used in the parametric study. For
comparisons with previous approaches, the CRes, the
CRot, and the stresses were also calculated with the tra-
ditional “no-fiber” (NPFM only) simplification. In
addition, the center locations were determined with
only fiber support (the no-matrix, or no-NPFM assump-
tion).
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 120, Number 3
B ment. A nonlinear static displacement/stress-analysis
Fig 6. Locations of CRes and CRot as functions of (A)
oblique fiber angulation (with fiber Young’s modulus
equal to 30 MPa) and (B) fiber stiffness (with fiber angu-
lation equal to 45°). The no-fiber case is represented by
the open symbols at 30° in A and at Young’s modulus
equal to zero in B. Similarly, the no-matrix (NPFM) con-
figuration is depicted as open symbols at 45° and
Young’s modulus is equal to 30 MPa in A and B, respec-
tively.
FE model and analysis
From the measured geometry and material selec-
tion, an FE model was generated (Figs 2 and 3) with
MSC/PATRAN 7.0 (MacNeal-Schwendler Corpora-
tion, Los Angeles, Calif). The model contains 16,320
eight-node elements and 18,216 nodes. The PDL ele-
ments are reinforced with (principal) fibers using spe-
cial ABAQUS 5.6 (Hibbitt, Karlson & Sorensen, Paw-
tucket, RI) elements in which the fibers are represented
in 1 of 2 ways. In the “single” type, the fiber connects
the centers of 2 opposing element faces. These ele-
ments are used in the representation of the apical PDL
Qian, Chen, and Katona 275
Fig 7. Locations of CRes and CRot. Solid circles indicate
5 combinations of 30°, 45°, and 60° fiber angulations
with 10, 30, and 90 MPa fiber Young’s moduli listed in
Table III. Open circles and squares represent no-fiber
(NPFM) and no-matrix configurations, respectively.
(Fig 4). For PDL not at the apex, “skew”-type elements
are applied (Fig 5).
As a boundary condition, all nodes on the 3 cut sur-
faces of the mandible were fixed. The mesial orthodon-
tic force, added on the lingual and buccal sides of the
crown (0.68 N × 2), was used in a laboratory experi-
process was used because the soft PDL matrix can
experience large strains and because the PDL fibers are
only allowed to transmit tension. Convergence was
assumed, because the largest displacement magnitude
changed by only 6%, with a 30% reduction in the num-
ber of elements.
RESULTS
The CRes and the CRot
As principal oblique fiber angulation decreased
from 60° to 30° (Young’s modulus = 30 MPa), the
CRes shifted away from the apex by about 0.13 mm
(Fig 6, A). Without fibers, it shifted an additional 0.52
mm relative to the position that corresponds to the 30°
fiber angulation. At the same time, the CRot shifted
toward the root tip by about 0.16 mm. Without fibers,
the CRot was about 1.3 mm farther away from the
apex.
With the Young’s modulus decreased from 90 to 10
MPa (fiber angulation = 45°), the CRes moved
occlusally by about 0.1 mm (Fig 6, B). At the same
time, the CRot moved about 0.93 mm farther from the
apex. Without fibers (fiber Young’s modulus = 0 MPa),
276 Qian, Chen, and Katona American Journal of Orthodontics and Dentofacial Orthopedics
September 2001

B
Fig 8. MAX in (A) lamina dura and (B) PDL NPFM change with oblique fiber angulation (30°, 45°,
60°, no-fiber) along path from distal coronal to apical to mesial coronal.

the CRes and the CRot moved away from the apex by
an additional 0.65 mm and 0.87 mm, respectively.
Calculations were performed with only fiber
(Young’s modulus = 30 MPa and angulation = 45°) sup-
port to help ascertain the relative influence of the
NPFM. These results are also depicted in Fig 6, A and
B. The CRot is displaced by only 0.12 mm apically, but
the CRes is 3.52 mm closer to the root tip. The results
presented in Fig 6, A and B, are combined in Fig 7.
Stress distributions in bone and NPFM
The stress parameters we calculated were the
maximum principal stress (MAX), the minimum
principal stress (MIN), and the maximum shear stress
(SHR) in the NPFM and the lamina dura. The results
from the distal and mesial sides of the root (Fig 2) are
displayed because they have the highest magni-
tudes.13
With the fiber Young’s modulus fixed (30 MPa),
stresses were calculated for 30°, 45°, and 60° oblique
fiber angulations (ie, f/L = 0.70, 0.47, and 0.13 in Fig
5) and without fibers. The values of MAX in the distal
and mesial regions of the lamina dura and the NPFM
are shown in Fig 8, A and B, respectively. These graphs
show that the principal oblique fiber angulation does
not change the general overall stress patterns in the
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 120, Number 3
B
Fig 9. As in Fig 8, except with varied fiber Young’s modulus (E = 90, 30, and 10 MPa, and no fiber).
PDL and the bone, but it does change the stress magni-
tudes. Similar effects are exhibited by the MIN and the
SHR.
Fiber orientation was defined as 45° to study the
influence of the principal PDL fiber Young’s modulus.
For the no-fiber model and fibers with Young’s modulus
10, 30, and 90 MPa, the MAX in the distal and mesial
regions of the lamina dura (Fig 9, A) and the NPFM (Fig
9, B) was calculated. In general, fiber Young’s modulus
does not seem to have much qualitative impact on the
stress patterns. Similar behavior is exhibited by MIN
and SHR.
Qian, Chen, and Katona 277
DISCUSSION
The ability to characterize, calculate, and predict
functional and orthodontic tooth movements and the
concomitant stresses within the alveolar bone and the
PDL are important in clinical dentistry and bone
research. The parametric analyses indicate that prin-
cipal-fiber angulation and stiffness influence the
tooth displacement and the mechanical environment
within the structures. Unfortunately, little is known
about the specific attributes of the individual fibers;
these are essential for the computations. The
unknowns include the fiber angulations, the angula-
278 Qian, Chen, and Katona
tion distributions and changes, and the mechanical
properties of the fibers.
Fortunately, these fiber characteristics appear to
have minimal, probably clinically insignificant, effects
on some calculations. For instance, oblique fiber angu-
lation only slightly changed the locations of the CRes
and the CRot (Fig 6, A), and fiber stiffness had a small
effect on the CRes (Fig 6, B). On the other hand, the 9-
fold difference in stiffness moved the CRot about 1.1
mm (Fig 6, B); that may be clinically relevant.
The overall stress patterns in the bone and the PDL
were essentially unaltered by changes in fiber angula-
tion and stiffness (Figs 8 and 9). In some studies, peak
stress component locations were linked to root resorp-
tion and bone resorption/formation patterns.5,14 In these
analyses, accurate knowledge of fiber properties may
not be necessary because it appears that peak locations
are virtually independent of fiber angulation and stiff-
ness.
It is significant, and perhaps not surprising, that the
traditional (ie, no-fiber) configuration behaves as an
“outlier.” The absence of principal fibers has a more
profound influence on the locations of the CRes and the
CRot than does doubling the fiber angulation (Fig 6, A)
or a 9-fold change in fiber stiffness (Fig 6, B). This is
illustrated in Fig 7 by the relatively tightly clustered
solid circles (the different combinations of fiber stiff-
nesses and angulations) and the open circles (the no-
fiber configuration).
The no-matrix (ie, only fibers in the PDL, no
NPFM) results (the open symbols at 45° in Fig 6, A; at
30 MPa in Fig 6, B; and the open squares in Fig 7) have
interesting implications. As displayed in Fig 7, the data
suggested that the location of the CRot is largely influ-
enced by the principal fibers, whereas the CRes is pri-
marily determined by the PDL matrix (NPFM). The
CRot and the CRes in the no-NPFM configuration were
separated by only 0.3 mm, compared with 3.0 mm for
the no-fiber case and 3.2 to 4.1 mm for the 5 models
that contain the matrix and the fibers. When considered
from this perspective, the no-fiber case and certainly the
no-matrix configuration behaved as “outliers.”
The MAX pattern in the PDL was the same with or
without fibers; however, the lack of fibers was associ-
ated with a relatively large quantitative difference (Fig
8, B, and Fig 9, B). The distribution of MAX in bone
was more affected by fiber angulation (Fig 8, A) and
stiffness (Fig 9, A) than was the MAX in the PDL, but
the largest local peak magnitudes of the MAX were
associated with the no-fiber case. An exception was the
distoapical region, in which the no-fiber case did not
predict a local peak.
American Journal of Orthodontics and Dentofacial Orthopedics
September 2001
Thus, qualitatively, principal-fiber orientation and
stiffness did not appear to play critical roles in the
mechanics of orthodontic tooth movement. Quantita-
tively, however, there were clear fiber angulation and
stiffness effects. The no-fiber configuration accentuated
the quantitative differences.
The fibers’ roles in other types of orthodontic tooth
movement (eg, lateral translation, intrusion, extrusion,
and axial rotation) and retention should also be evalu-
ated. Furthermore, it is conceivable that some nuances,
such as the absence of the no-fiber peak in bone in the
distoapical region (Fig 8, A, and Fig 9, A), may be
related to the specific anatomy of this root.
CONCLUSION
We have shown that, for quantitative studies of
tooth movement, the principal PDL fibers should be
included. Unfortunately, the necessary data about
fiber orientations, distributions, and mechanical prop-
erties, and how they change during tooth movement
and function, are not available. However, the calcula-
tions (Figs 6-9) suggested that the results for the var-
ious combinations of fiber orientation and stiffness
are more similar to each other than to the results from
the model without any fibers (Fig 7). Thus, it may be
more prudent to incorporate what may be unrealistic
principal fibers into FE models than not to include
them at all.
We gratefully acknowledge the assistance of Dr Tod
Curtis.
REFERENCES
1. Chen J, Katona TR. The limitations of the instantaneous centre of
rotation in joint research. J Oral Rehabil 1999;26:274-9.
2. Middleton J, Jones M, Wilson A. The role of the periodontal lig-
ament in bone modeling: the initial development of time-depen-
dent finite element model. Am J Orthod Dentofacial Orthop
1996;109:155-62.
3. Puente MI, Galban L, Cobo JM. Initial stress differences between
tipping and torque movements: a three-dimensional finite ele-
ment analysis. Eur J Orthod 1996;18:329-39.
4. Tamatsu Y, Kaimoto K, Arai M, Ide Y. Properties of the elastic
modulus from buccal compact bone of human mandible. Bull
Tokyo Dent Coll 1996;37:93-101.
5. Katona TR, Paydar NH, Akay HU, Roberts WE. Stress analysis
of bone modeling response to rat molar orthodontics. J Biomech
1995;28:27-38.
6. Tanne K, Sakuda M, Burstone CJ. Three-dimensional finite ele-
ment analysis for stress in the periodontal tissue by orthodontic
forces. Am J Orthod Dentofacial Orthop 1987;92:499-505.
7. Auyeung L, Bouwsma OJ, Polson AM. Periodontal fiber attach-
ment and apical root resorption. Endod Dent Traumatol
1988;4:219-25.
8. Kurihara S, Enlow DH. An electron microscopic study of attach-
ments between periodontal fibers and bone during alveolar
remodeling. Am J Orthod 1980;77:516-31.
American Journal of Orthodontics and Dentofacial Orthopedics Qian, Chen, and Katona 279
Volume 120, Number 3

9. Bernick S. The organization of the periodontal membrane fibres
of the developing molars of rats. Arch Oral Biol 1960;2:57-63.
10. Atmaram GH, Mohammed H. Estimation of physiologic stresses
with a natural tooth considering fibrous PDL structure. J Dent
Res 1981;60:873-7.
11. Katona TR, Qian H. A mechanism of non-continuous supra-
osseous tooth eruption. Am J Orthod Dentofacial Orthop 2001; In
press.
12. Ralph WJ. Tensile behaviour of the periodontal ligament. J Peri-
odontal Res 1982;17:423-6.
13. Gianelly AA, Goldman HM. Biological basis of orthodontics.
Philadelphia: Lea & Febiger; 1971.
14. Katona TR. A mechanical engineering analysis of orthodontics-
associated external root resorption. In: Davidovitch Z, editor. The
biological mechanisms of tooth eruption, resorption and replace-
ment by implants. Birmingham (AL): EBSCO Media; 1994. p.
537-43.
15. O’Grady J, Sheriff M, Likeman P. A finite element analysis of a
mandibular canine as a denture abutment. Eur J Prosthodont
Restor Dent 1996;4:117-21.

CORRECTION

Table I in “Effects of combined application of antimicrobial and fluoride varnishes in orthodontic patients” (Am
J Orthod Dentofacial Orthop 2001;120:28-35) was published incorrectly. The corrected table is given below.

Table I. Percentage of specified teeth with white spot lesions at debonding

Tooth (FDI tooth number)

16 15 14 13 12 11 21 22 23 24 25 26
Group 1 14.7 7.6 14.3 13.6 12.3 1.8 0.9 11.3 11.8 7.8 3.9 18.4
Group 2 16.4 6.1 7.6 15.5 19.2 5.5 4.6 20.1 11.0 14.8 6.0 15.5
Group 3 20.2 14.9 27.3 26.0 27.4 9.0 13.0 31.9 25.0 20.5 12.5 23.0
Tooth (FDI tooth number)
46 45 44 43 42 41 31 32 33 34 35 36
Group 1 29.4 13.6 21.3 12.7 8.3 4.7 2.8 0.0 15.5 25.3 20.7 38.0
Group 2 35.8 24.1 35.4 17.3 9.1 11.7 6.4 11.8 14.6 25.0 23.6 33.6
Group 3 40.9 26.1 31.3 18.8 10.4 6.7 5.7 12.2 17.8 27.8 26.9 40.0