Periodontal force: A potential cause of relapse

Thomas E. Southard, DDS, MS," Karin A. Southard, DDS, MS, b and Elizabeth A. Tolley, PhD
Iowa City, Iowa, and Memphis, Tenn.

Relapse of aligned mandibular anterior teeth and the progressive collapse of the mandibular arch is
a significant problem for orthodontists. However, identification of a specific cause of such relapse
has proved elusive. The transseptal fiber system is thought to stabilize teeth against separating
forces. It is hypothesized that this fiber system may actually maintain the contacts of approximating
teeth in a state of compression, the long-term result of which could be contact slippage and collapse
of the arch. The interproximal force (IPF) at the mandibular first molar-second premolar contact was
investigated on the basis of previous studies with this representative contact. The IPF was measured
in 10 subjects at six different widths of contact separation. By means of graphic plotting techniques,
the IPF at zero separation was calculated to estimate the contact force when the molar and premolar
were actually touching. The mean IPF at zero separation was found to be 36.7g (SE = 6.6g), and a
t test confirmed the hypothesis that a state of compression between contacts exists (p < 0.0001).
After chewing, the mean IPF at zero separation was 57.2g (SE = 9.1g), and a paired t test revealed
an increase in contact compression had resulted (p < 0.01). It was concluded that the periodontium
exerts a continuous force on the mandibular dentition and that this force acts to maintain the
contacts of approximating teeth in a state of compression. This force is increased after occlusal
loading and may help to explain long-term crowding of the mandibular anterior teeth, physiologic
drifting of teeth, and maintenance of posterior dental contacts after interproximal wear. (AM J
ORTHOD DENTOFACORTHOP 1992;1 01:221-7.)

P e r h a p s one of the most striking deficiencies
in orthodontics is our lack of understanding of post-
treatment relapse. Dentists claim that almost 50% of
the United States population needs orthodontic treat-
ment, 1 but 90% of the patients treated have unaccept-
able dental alignment 10 to 20 years after retention. 2
Although orthodontic appliances and mechanics used
to correct malocclusions have improved in recent
decades, the identification of etiologic factors that cause
corrected alignment to relapse has proved elusive.
Assessment of treated orthodontic patients has
shown no descriptive characteristics (such as Angle
class, length of retention, age at start of treatment, or
gender), no dental model parameters (such as initial or
end-of-treatment alignment, overbite, overjet, arch
width, or arch length), and no cephalometric parameters
(such as maxillary and mandibular incisor proclination,
patterns of horizontal and vertical growth, or mandib-
ular plane angle) that are of value in predicting long-
term stability of dental alignment. 2'3 The recent rec-
ommendation that lifelong retention be used to ensure
aAssistant Professor, Department of Orthodontics, The University of Iowa.
bAssistant Professor, Department of Orthodontics, The University of Iowa.
CAssistant Professor, Department of Biostatistics and Epidemiology, The Uni-
versity of Tennessee.
8/1/26754
posttreatment dental alignment 2 clearly underscores our
lack of understanding of relapse etiology.
The periodontium itself may be linked to posttreat-
ment relapse. As cited 50 years ago by Waldron, 4 Ru-
dolf Kronfeld suggested that a function of the trans-
septal fiber system was to stabilize the teeth against
separating forces. If such stabilization is accomplished
by maintaining the contacts of approximating teeth in
a slight state of compression, then the long-term effect
of this compression could be slippage of dental contacts
and collapse of the arch. The purpose of this study was
to determine whether the periodontium maintains the
contacts of approximating mandibular teeth in a con-
tinuous state of compression.
EXPERIMENTAL HYPOTHESIS
If a thin material is inserted between two approxi-
mating teeth, the two teeth will be displaced mesio-
distally by a distance (x), the width of the inserted
material. Therefore, the displaced teeth will exert a
reactive force against the material, a force that we define
here as the interproximal force (IPF). Intuitively, IPF
should vary as a function of the material thickness. That
is,
IPF = IPF(x) or IPFx (1)
An estimate of IPFo, the actual contact force that

221
222 Southard, Southard, and Tolley
lil,5 ~
DIGITAL TENSION TRANSDUCER
Fig. 1. Interproximal force (IPF) measurement technique. A, The tension transducer was hooked
through a perforated stainless steel strip. B, The strip was slipped interproximally and withdrawn.
exists between approximating teeth with no material
inserted (x = 0.0 mm), may be estimated by measuring
IPFx with .different material thicknesses, graphically
plotting the relationship between IPFx and x, determin-
ing a suitable equation to approximate this relationship,
and then extrapolating IPFx to x = 0.0 mm. On the
basis of the work done by Parfitt 5 and our own under-
standing of biomechanics, we hypothesized that IPFx
varies exponentially as a function of x and that
I P F o > 0.0g.
MATERIALS AND METHODS
Technique of measuring IPFx
The technique used to measure IPFx was previously dis-
cussed and was based on frictional force concepts. 6'7 Briefly,
a thin stainless steel strip (6 x 15 mm) was slipped into the
mandibular left first molar-second premolar contact and was
withdrawn with a digital tension transducer (Fig. 1). IPG
was related to the frictional force resisting withdrawal (FF)
and the coefficient of dynamic friction (Ix) between tooth
enamel and stainless steel by the following equation:
IPF~ = G/21x (2)
Force measurements were made with strips of six
different thicknesses. In other words, IPFx was mea-
sured for six values of x. These thicknesses were x =
0.00279 mm (Hamilton Precision Metals, Lancaster,
Penn.), 0.00764 m m (Hamilton Precision Metals, Lan-
caster, Penn.), 0.0153 mm, 0.0229 mm, 0.0305 mm,
and 0.0367 m m (Healthco, Inc., Boston, Mass.). The
0.0153 mm, 0.0229 mm, and 0.0305 mm strips were
fabricated with layers of steel 0.00764 mm thick.
Since the three steels had slightly different surface
Am. J. Orthod. Dentofac. Orthop.
March 1992
finishes, we constructed an apparatus to determine the
value of p~ for each steel. A known IPFx was applied
between sectioned halves from five extracted teeth. A
strip made from a selected steel was slipped interprox-
imally, and FF was measured as the strip was with-
drawn. The mean coefficient of dynamic friction was
calculated for the five extracted teeth with equation 2
and found to be 0.178 (SE = 0.007), 0.161 (SE =
0.004), and 0.137 (SE = 0.001) for strips of 0.00279
mm, 0.00764 mm, and 0.0367 m m thickness, respec-
tively. Although we attempted to measure IPFx through-
out our experiments in a dry field, saliva contamination
was possible. We therefore decided to find the coeffi-
cient of dynamic friction in the presence of saliva for
each steel and to average this value with that previously
found under dry conditions to give an overall value of
~z for each steel. The coefficients of dynamic fric-
tion in the presence of saliva were found to be
0.209 (SE = 0.012), 0.171 (SE = 0.009), and 0.139
(SE = 0.004), and the average dry and wet values were
found to be 0.193, 0.166, and 0.138 for steel strips of
0.00279 m m , 0.00764 mm, and 0.0367 mm thickness,
respectively.
Two digital force gauges were used. The most sen-
sitive gauge (Accuforce Cadet Force Gauge, Ametek,
Largo, Fla.) had an accuracy of 0. l g and was used for
measurements with the three thinnest strips. A less sen-
sitive gauge (Accuforce III, Ametek, Largo, Fla.), with
an accuracy of 1.0g, was used for larger measurements
with the three thicker strips.
We chose to study the left first molar-second premolar
contact as a representative mandibular contact for two rea-
sons. First, access precluded measurement of the second
 Volume 101 -- - - ojvPerlo'~rce
~lonta/ 223
Number 3

molar-first molar contact. Second, we had previously dem- IPF V A R I A T I O N - SUBJECT 1 INITIAL
onstrated that the first molar-second premolar contact has the 10000
highest IPF magnitude of all contacts in either arch when
measured with a 0.0367 mm steel strip? The thinnest steel
strips that we used in this study were approximately 1/15 of 1000
the 0.0367 mm strip thickness, and we anticipated that the
force readings with the thinnest strips would be small. It
therefore seemed reasonable to study the force in the contact v
100
that would provide the largest possible readings for each strip 14.
13.
m
thickness. That is, if IPFo > 0.0g than our ability to detect,
it would be greatest at the tightest contact. Finally, the left 10
side of the arch was chosen because of operator convenience.

Subject selection and experimental procedure

1 i i i
Ten men participated. All the subjects had complete den- 0.00 0.01 0.02 0.03 0.04

titions from the second molars forward, had approximating A

contacts between all adjacent mandibular teeth, and had been x [mm~
requested not to eat for at least 1 hour before the experiment.
Each subject rinsed his mouth with water, and the mandibular
IPF V A R I A T I O N - S U B J E C T 1 AFTER CHEWING
first molar-second premolar contact was air dried. With the
maxillary and the mandibular dental arches apart, frictional 10000
forces were first measured with the thinnest stainless steel
strip and then repeated with progressively larger strips.
IPF x = 5 2 . 7 e 86"9x R 2 = 0.99
Each subject then chewed bilaterally on hard crackers for 1000
5 minutes. Frictional forces were measured with the thinnest
strip. The measurement procedure was then repeated with
progressively thicker strips after 1-minute intervals of chew- 100 '
14.
ing. All values of IPF~ were then calculated, both initially it.
and after chewing, with equation 2.
10
Statistical analysis
Statistical analyses were performed with the SAS 9 statis-
tical software package. The nonlinear regression procedure 1 i i i
0.00 0.01 0.02 0.03 0.04
produced estimates of the parameters IPFo and k for the fol-
lowing model: B
x (mm)
IPFx = IPFoe~ (3)
Fig. 2. Mandibular first molar-second premolar interproximal
This model is a transformation of the relationship between force, measured with six different widths (x) of stainless steel,
displacement of teeth and applied force as described by for a typical subject. A, Initial measurements and best-fit model
Parfitt? The model simply states that the interproximal force representing this relationship on a semilog plot. Note that the
(IPFx), measured for any mesiodistal displacement of teeth estimated IPFx intercept, at x = 0.0 mm (IPFo), is 19.4g. B,
(x), equals the interproximal contact force at zero mesiodistal Measurements after chewing. Note that IPFo has increased from
19.4g to 52.7g after chewing.
displacement (IPFo) multiplied by an exponential function
(ekx). The exponent is a product of the displacement multiplied
by some constant, k, which should vary between subjects.
Note that at a displacement of 0.0 mm,
A t test was used to test our hypothesis that IPFo > 0.0g. A
IPFx = IPFoe"x = IPF,,e = IPFo (4) paired t test was also used to detect changes in IPFo after
chewing.
The IPFo and k coefficients were generated for each sub-
ject twice, initially and after chewing. Experimental values RESULTS
were weighted as the inverse of the standard error of the mean The m a n d i b u l a r first m o l a r - s e c o n d p r e m o l a r inter-
to correct for increasing IPFx variance with increasing x (het- p r o x i m a l force, measured with six different widths (x)
eroscedasticity). R 2, the proportion of variation accounted for of stainless steel, is plotted in Fig. 2, A for a typical
by the model, was calculated for each model as:
subject. T h e mathematical m o d e l , generated by the non-
R2 = (Regression sum of squares) (5) linear regression procedure, w h i c h closely fits this re-
(Uncorrected total sum of squares) lationship (R 2 = 0.99), is:
224 Southard, Southard, and Tnlle,~ Am. J. Orthod. Dentofac. Orthop.
March 1992

Table I. Interproximal force (IPFx) model* coefficients for each subject, generated from
initial measurements
IPFo asymptotic 95%
confidence interval

Subject IPFo Lower Upper Model R2

1 19.4 7.6 31.1 112.7 0.99

2 86.9 40.5 133.3 97.4 0.99
3 45.8 23.5 68.1 89.1 0.98
4 6.0 -4.8 16.9 102.8 0.90
5 29.7 22.4 37.0 103.3 0.99
6 34.9 19.1 50.7 88.2 0.98
7 33.0 - 3.3 69.3 103.1 0.96
8 37.1 9.1 65.1 82.0 0.96
9 33.9 23.4 44.4 98.5 1.00
10 40.5 16.6 64.4 106.4 0.99

*IPFx = IPFoekX[g].

Table II. Interproximal force (IPFx) model* coefficients for each subject, generated from measurements
after chewing
IPFo asymptotic 95%
confidence interval

Subject IPFo Lower Upper Model R z

1 52.7 36.4 68.9 86.9 0.99

2 ' 127.6 -36.9 292.3 87.4 0.93
3 84.6 -6.6 175.9 72.3 0.91
4 51.6 8.0 95.1 68.9 0.94
5 41.3 32.7 49.8 95.1 1.00
6 43.5 19.9 67.0 83.1 0.98
7 59.8 -6.0 125.7 94.4 0.98
8 46.3 9.0 83.6 81.3 0.96
9 31.8 11.9 51.7 107.4 0.99
10 32.6 12.3 53.0 113.7 0.99

*IPFx = IPFoee' [g].

IPFx = 19.4e .... [g] (6)

In other words, for this subject an estimated 19.4g
(IPFo) of compressive force is present when the first
molar and the second premolar are actually contacting
(x = 0.0 mm). Fig. 2, B plots the variation of inter-
proximal force with steel strip width after chewing for
this same subject. The mathematical model describing
this relationship is as follows:
IPFx = 52.7e ..... [g] (7)
That is, after chewing, the estimated compressive force
existing between these teeth when they are actually
contacting was more than doubled.
IPFo and k model coefficients for all 10 subjects,
generated from measurements made initially and after
chewing, are presented in Tables I and II, respectively.
Continuous compressive forces were estimated to exist
between the teeth in all subjects. Initial IPFo ranged
from 6.0g to 86.9g. The smallest initial IPFo was noted
in subject 4, who reported not having chewed anything
during the previous 20 hours. The mean initial IPFo was
36.7g (SE = 6.6g). The mean asymptotic 95% con-
fidence interval for initial IPFo was 15.4g to 58.0g.
IPFo increased in eight subjects and decreased in
two subjects after chewing. Its value after chewing
ranged from 31.8g to 127.6g. The mean IPFo after
chewing increased to 57.2g (SE -- 9. lg), and the mean
asymptotic 95% confidence interval ranged from 8.0g
to 106.0g. It should be noted that subject 4 also had
the greatest increase in IPFo (800%) after experimental
chewing.
A plot o f the mean initial variation of interproximal
 Volume 101
Number 3 Periodontal jorce 225

force with steel strip width is present in Fig. 3, A, and IPF V A R I A T I O N - MEAN INITIAL
a plot of the same relationship after chewing is provided 10000
in Fig. 3, B. As a result of the t tests, our original
hypothesis that initial IPFo > 0.0g was determined to IPF x = 36,7e 98-4x R 2 = 0.97 ..~ "~
be correct (p < 0.0001) and an increase in IPFo after 1000

chewing was demonstrated (p < 0.01).

DISCUSSION 100
I.[.
D,
The results of our research indicate that contacts of m

approximating mandibular teeth are maintained in a 10

continuous state of compression. This compressive
force is generated by the supporting periodontium and
acts through the dental contact points, even when the 1 I I I

dental arches are apart. Further, this force is increased 0.00 0.01 0.02 0.03 0.04

for a period after chewing. A x (mm)

The existence of IPFo could be of clinical impor-
tance in explaining several phenomena, such as long-
term collapse of the mandibular dental arch and mi- IPF V A R I A T I O N - MEAN A F T E R C H E W I N G
gration of teeth. However, as stated by Edwards, l the
10000
physiologic basis for the periodontal generation of a
force is an enigma. In a study of approximal tooth drift
IPF x = 57.2e 89.1x R 2 = 0.96
in monkeys, Moss and Picton ]~ proposed that removal 1000
of approximal tooth contact allowed the transeptal fiber
system to contract and produce approximation of the
adjacent teeth. Such fiber contracture could account for 100
I1.
IPFo. Another possible explanation is that teeth are com- a.

pressed and held mesially under action of the anterior

10
component of occlusal force. 6 This effect could also
explain the increase in mean IPFo after chewing that
we observed. 1 I I I

The major limitation of this research was the use
of curve fitting and extrapolation to obtain IPFo. An
exponential equation appears suitable for approximat-
ing the graphic relationship between IPF~ and x (Figs.
2 and 3), but there always exists a danger in extra-
polating a function to an area outside the given data
range, in this case to x = 0.0 mm. However, as pre-
viously noted, we have greater than 97.5% confidence
that the true mean IPFo > 15.0g. Perhaps an IPFo of
much less magnitude could be capable of exerting pro-
nounced clinical effects since teeth can be moved by a
continuously acting force of only a few grams.'2
Another limitation was that IPFo was estimated only
for the mandibular first molar-second premolar con-
tact. However, our findings should apply to all man-
dibular approximating contacts, although we anticipate
that the respective IPFo magnitudes will be less.
Posttreatment relapse and clinical implications
Some mechanism in modern man causes a progres-
sive long-term collapse of the mandibular anterior den-
tal arch. Decreased intercanine width, decreased arch
length, and increased incisor crowding are seen over
0.00 0.01 0.02 0.03 0.04
B x (mm)
Fig. 3. Mean (n = 10) mandibular first molar-second premolar
interproximal force, measured with six different widths (x) of
stainless steel. The error bars indicate the standard error of the
mean. A, Initial measurements and best-fit model on a semilog
plot. B, Measurements and best-fit model after chewing.
time in persons with untreated normal occlusions and
in persons receiving orthodontic treatment with or with-
out extractions. 2.13.14Relapse of orthodontically aligned
anterior teeth is unpredictable; no cephalometric param-
eters, including growth, are useful in establishing a
prognosis. 3
The existence of a continuous, compressive force
(IPFo), originating in the periodontium and acting on
approximating teeth at their contact points, may help
explain long-term arch collapse (Fig. 4). Although we
have previously found correlations (0.53 to 0.59,
p < 0.05) between mandibular anterior malalignment
and IPF measured with the 0.0367 mm thick steel
 *~,, JT~'te"
226 Southard, Southard, and
Fig. 4. A, Schematic representation of continuous compressive
interproximal contact force (iPFo) originating in periodontium
and acting through dental contacts. B, The action of such a
force may help explain long-term slippage of dental contacts
and collapse of the mandibular anterior arch.
strip, 1~it would be simplistic to suggest that IPFo could
be solely responsible for arch collapse. Rather, if IPFo
does exert an influence on dental alignment, it probably
acts in conjunction with lip and cheek forces to collapse
the arch and is opposed by tongue force, which tends
to expand the arch. It follows that the influence of IPFo
should be more evident in the anterior region of the
arch where the contact points are narrower, the crowns
more tapered, and the expansive force from the tongue
more intermittent than in the posterior region of the
arch.
The IPFo varies widely between persons and be-
tween periods of occlusal function. If IPFo is a factor
in posttreatment crowding, then its influence also varies
dramatically between persons and probably acts as a
function of their occlusal loading, including parafunc-
tional habits.
When a rapid maxillary expansion appliance is used
to open the midpalatal suture, the orthodontist antici-
pates both the formation of a diastema between the
Am. J. Orthod. Dentofac. Orthop.
M a r c h 1992
central incisors and its rapid, spontaneous closure (re-
lapse). Such closure is based on rebound of the stretched
periodontium and not on lip and tongue forces. Simi-
larly, we speculate that arch expansion and excessive
incisor proclination may stretch the periodontium, pos-
sibly resulting in increased IPFo levels, and may ac-
celerate arch collapse after retention. In other words,
a strict nonextraction approach to orthodontic treatment
of dental crowding, irrespective of the resulting degree
of arch expansion, could be questioned.
Attrition, migration, and equilibrium positions of
the teeth
The proximal contact of teeth is maintained under
normal circumstances in spite of continual interproxi-
mal wear and the resultant decrease in mesiodistal
crown dimension. The existence of a force that contin-
uously compresses tooth crowns together may help ex-
plain this observation because as teeth wear interprox-
imally, they may be held together by IPFo. Our finding
that IPFo increases with chewing may also explain the
high levels of mandibular mesiodistal attrition found by
Begg 16 in the Australian aboriginal man. That is, with
a tough aboriginal diet, greatly increased IPFo forces
would be anticipated during chewing and increased
enamel abrasion would result as the more compressed
teeth move across each other at the interdental contacts.
After extraction of first premolars in crowded man-
dibular arches, the canines drift laterally and distally
into the extraction sites, whereas the incisors become
more upright over basal bone and less crowded. 17
Alexander 18has suggested delayed banding of the lower
arch after extraction to allow unaided physiologic distal
drifting and unraveling of the anterior teeth. Such spon-
taneous migration through an extraction site could be
evidence of the same inherent periodontal force that
causes IPFo.
Occlusal forces have been dismissed as having too
short a duration to move teeth buccally or lingually.12
However, the results of this study indicate that occlusal
loading of teeth causes them to become increasingly
compressed together, at least in the "contained" man-
dibular arch, for a period after the occlusal load is
withdrawn. That is, occlusal forces exhibit a prolonged
influence and may have a significant effect in deter-
mining the equilibrium position of teeth.
In conclusion, the results of this study indicate that
the periodontium exerts a continuous force on the man-
dibular dentition, and this force acts to maintain con-
tacts of approximating teeth in a state of compression.
This force is increased after occlusal loading and may
Volume 101
Number 3 Periodontal Jorce 227

help to explain posttreatment collapse of the mandibular
arch, physiologic drifting, and contact maintenance in
the presence of interproximal wear.
REFERENCES
l. Bell WM, Proffit WR, White RP. Introduction. In: Bell WM,
Proffit WR, White RP, eds. Surgical correction of dentofacial
deformities. Philadelphia: WB Saunders, 1980:3-4.
2. Little RM, Riede] RA, Artun J. An evaluation of changes in
mandibular anterior alignment from 10 to 20 years postretention.
AM J ORTHODDENTOFACORTHOP 1988;93:423-8.
3. Shields TE, Little RM, Chapko MK. Stability and relapse of
mandibular anterior alignment: a eephalometric appraisal of first
premolar extraction cases treated by traditional edgewise ortho-
dontics. AM J ORTHOD 1985;87:27-38.
4. Waldron R. Reviewing the problem of retention. AM J ORTHOD
ORAL SURG 1942;28:770-91.
5. Parfitt GJ. The dynamics of a tooth in function. J Periodontol
1961 ;6:102-6.
6. Southard TE, Behrents RG, Tolley EA. The anterior component
of occlusal force. Part 1. Measurement and distribution. AM J
ORTHOD DENTOFACORTHOP 1989;96:493-500.
7. Osborn JW. An investigation into the interdental forces occurring
between the teeth of the same arch during clenching the jaws.
Amh Oral Biol 1961;5:202-11.
8. Southard TE, Southard KA, Tolley EA. Variation of approximal
tooth contact tightness with postural change. J Dent Res
1990;69:1776-9.
9. SAS Institute, Inc. SAS user's guide: basics, 5 ed. Cary, N.C.:
SAS Institute, 1985.
10. Edwards JG. A long-term prospective evaluation of the circum-
ferential supracrestal fiberotomy in evaluating orthodontic re-
lapse. AM J ORTHOD DENTOFACORTHOP 1988;93:380-7.
11. Moss JP, Picton DCA. Short-term changes in the mesiodistal
position of teeth following removal of approximal contacts in
the monkey Macaca fascicularis. Arch Oral Biol 1982;27:273-
8.
12. Proffit WR. Equilibrium theory revisited: factors influencing po-
sition of the teeth. AM J ORTHOD 1978;48:175-86.
13. Ades AG, Joondeph DR, Little RM, Chapko MK. A long-term
study of the relationship of third molars to changes in the man-
dibular dental arch. AM J ORTHOD DENTOFAC ORTHOP 1990;
97:323-35.
14. Sinclair PM, Little RM. Maturation of untreated normal occlu-
sions. AM J ORTHOD 1983;83:114-23.
15. Southard TE, Behrents RG, Tolley EA. The anterior component
of occlusal force. Part 2. Relationship with dental malalignment.
AM J ORTHOD DENTOFACORTHOP 1990;97:41-4.
16. Begg PR. Stone Age man's dentition. AM J ORTHOD 1954;
40:298-312, 383, 462-75, 517-31.
17. Papandreas SB. An analysis of drift in the mandibular dentition
following first premolar extractions [Masters thesis]. Dallas,
Texas: Baylor University, 1989.
18. Alexander RG. The vari-simplex discipline. Part 3, Extraction
treatment. J Clin Orthod 1983;17:537-47.
Reprint requests to:
Dr. Thomas E. Southard
Department of Orthodontics
College of Dentistry
The University of Iowa
Iowa City, IA 52242