CONTINUING EDUCATION ARTICLE

Stability of orthodontic treatment outcome: Follow-up until 10

years postretention
Essam A. Al Yami, DDS,a Anne M. Kuijpers-Jagtman, DDS, PhD,b and Martin A. van t Hof, PhDc
Nijmegen, The Netherlands
Dental casts of 1016 patients were evaluated for the long-term treatment outcome using the Peer
Assessment Rating (PAR) index. The PAR index was measured at the pretreatment stage (n = 1016),
directly posttreatment (n = 783), postretention (n = 942), 2 years postretention (n = 781), 5 years
postretention (n = 821), and 10 years postretention (n = 564). The mean absolute change as well as the
percentage of change per year (relapse) related to the postretention stage was calculated. An analysis of
variance was applied to compare the mean change in the PAR between cases with and without a fixed
retainer at the postretention stage and up to 10 years postretention. Drop-out analysis showed that more
Class II Division 2 cases were lost to follow-up than cases of other Angle classes. The results indicate that
67% of the achieved orthodontic treatment result was maintained 10 years postretention. About half of the
total relapse (as measured with the PAR index) takes place in the first 2 years after retention. All occlusal
traits relapsed gradually over time but remained stable from 5 years postretention with the exception of the
lower anterior contact point displacement, which showed a fast and continuous increase even exceeding the
initial score. The presence of a fixed retainer had a positive effect on the PAR score. In cases with fixed
retention, the relapse was 3.6 PAR points less at 5 years postretention and 4.6 points less at 10 years
postretention. The results of this type of studies enable clinicians to inform their patients about treatment
limitations in order to better meet their expectations. (Am J Orthod Dentofacial Orthop 1999;115:300-4)

valuation of treatment results and longterm posttreatment assessment of orthodontically treated malocclusions has been of interest for several
decades.1-3 Follow-up studies of treated cases have
shown that although improvement in the dentition can
obviously be achieved, there is a tendency to return
toward the original malocclusion many years posttreatment.4-9 It is obvious that there is a large variability in
orthodontic treatment outcome for different individuals. This variability may be due to severity and type of
malocclusion, treatment approach, patient cooperation,
growth, and adaptability of the hard and soft tissues.
Additional factors that may influence the stability of
orthodontic treatment are the type, duration, and the
timing of the retention appliance.10
Most studies are concerned with the description of
long-term stability of specific treatment regimens for
aOrthodontist, Department of Orthodontics and Oral Biology, University of
Nijmegen, The Netherlands.
bProfessor and Head, Department of Orthodontics and Oral Biology, University of Nijmegen, The Netherlands.
cAssociate Professor, Department of Biostatistics and Epidemiology, University of Nijmegen, The Netherlands.
Reprint requests to: A. M. Kuijpers-Jagtman, DDS, PhD, Department of Orthodontics and Oral Biology, University of Nijmegen, PO Box 9101, 6500 HB
Nijmegen, The Netherlands; e-mail, orthodontics@dent.kun.nl
Copyright 1999 by the American Association of Orthodontists.
0889-5406/99/$8.00 + 0 8/1/91941

300

specific types of malocclusion such as the Angle Class

II/1.8,11-14 Other studies evaluated the stability of
orthodontic treatment outcome longitudinally for specific occlusal traits such as open bite, overbite, overjet,
posterior crossbite, intercanine and intermolar distance, and lower anterior crowding.12,13,15,16
In recent years the Peer Assessment Rating (PAR)
index was developed to assess treatment outcome in a
quantitative manner.17 The PAR index offers uniformity, objectivity, and standardization in assessing the outcome of orthodontic treatment. Also the index is
amenable to statistical analysis and easy to apply.17-20
In the literature, only two studies were found that
assessed the long-term stability of orthodontic treatment in a quantitative manner using the PAR
index.14,21 Fox and Chadwick21 found a PAR reduction
of 72% (from 29.5 to 8.3) in 100 cases at posttreatment. This reduction relapsed to 57% (12.8, n = 51) at
1 year postretention. Otuyemi and Jones14 evaluated 50
Class II/1 malocclusions. The results indicated a PAR
reduction of 82%. Maintenance of posttreatment
results at 1 and 10 years postretention was only
achieved in 60% and 38% of the cases, respectively.
The major factor involved in this deterioration
appeared to be late lower anterior crowding.
Kahl-Nieke et al22 emphasized that the review of the
literature points out the need for a quantitative and qual-

Al Yami, Kuijpers-Jagtman, and Van t Hof 301

American Journal of Orthodontics and Dentofacial Orthopedics

Volume 115, Number 3

Table I. Number

of patients, sex distribution, and mean

age ( SD) at all observational stages
Age

TP
T00
T0
T2
T5
T10

Total (n)

Male (n)

Female (n)

Mean

SD

1016
783
942
781
821
564

447
334
420
357
370
239

569
449
522
424
451
325

12.0
15.6
16.7
18.7
21.8
26.3

3.1
3.0
3.1
3.0
3.2
2.9

TP = Pretreatment; T00 = posttreatment; T0 = postretention;

T2 = 2 years postretention; T5 = 5 years postretention;
T10 = 10 years postretention.

Table II. Mean

PAR and mean absolute and percentage

change (SD) at all observational stages with respect
to pretreatment (TP). (For explanation of stages see
Table I.)
Stages
TP
T00
T0
T2
T5
T10

PAR

Absolute
change to TP

Percentage
change to TP

1016
783
942
781
821
564

28.4 10.2
8.5 6.7
9.5 7.3
12.2 8.4
13.6 8.9
14.6 9.7

20.3 11.4
19.1 11.2
16.3 11.5
14.8 11.7
13.7 12.0

67.1 27.1
63.8 28.8
54 32.4
48.7 33.7
45.2 36.3

itative assessment of posttreatment changes by using a

sample that is large enough for statistical analysis consisting of cases out of retention for at least 10 years. At
the University of Nijmegen, such a large treated sample
is available (more than 1000 cases). The aim of this study
was to evaluate the long-term posttreatment results until
10 years postretention using the PAR index.
MATERIAL AND METHODS
Subjects

The archives of the Department of Orthodontics

and Oral Biology, University of Nijmegen, contain
records of 2368 patients available for follow-up study.
In 1016 patients, treatment was started before the year
1982, which means that these patients were at least 10
years out of retention when the present study was performed (1995). Only the patients with at least a pretreatment and posttreatment or postretention dental cast
were included in this study. Patients with damaged dental casts or with prosthetic replacements that would
affect the measurements were excluded. Dental casts
were routinely made at the following stages: pretreatment (TP); posttreatment (T00); postretention (T0),
which means cessation of all removable retainers but

Table III. Mean

relapse in PAR between consecutive

stages. Mean percentage change is relative to the total
relapse over 10 years (5.61 PAR point), the retention
period (T00 to T0) is commonly 1 year. (For explanation of stages see Table I.)

Stages

Mean age
at start
of interval

T00-T0
T0-T2
T2-T5
T5-T10

709
745
658
497

15.5 2.7
16.5 2.9
18.5 3
21.4 2.9

Mean
PAR
change

Mean %
PAR
change

Mean %
PAR
change/year

0.56 6.4
2.75 5.5
1.57 4.2
0.73 4.0

10.0%
49.0%
28.0%
13.0%

10.0%
24.5%
9.3%
2.6%

bonded retainer wires may be maintained; 2 years after

T0 (T2); 5 years after T0 (T5) and 10 years after T0
(T10). Table I shows the number of patients, the sex
distribution, and the mean age ( standard deviation
[SD]) at all stages. More insight into the drop-out of
the study was obtained by comparing patients who had
no missing dental casts (n = 400) with patients who had
any missing cast at T2, T5, or T10.
A subdivision was made between patients with and
without a fixed retainer. Patients were divided into two
groups: cases with (n = 110) and without (n = 888)
fixed retainer. The 18 patients with unknown type of
retention were excluded. In comparing relapse in
patients with and without a retainer, the phenomenon
of confounding by indication may play a role. Possible confounders such as the initial PAR and PAR subscores were taken into consideration.
Methods

The PAR index18,19 was used to score pretreatment,

posttreatment, and all available postretention dental
casts of the same patient. The index has seven components: upper contact point displacement, lower contact
point displacement, left buccal occlusion, right buccal
occlusion, overjet, overbite, and centerline. The individual scores for the various components are finally
summed up for the so-called weighted PAR score, to be
called PAR in this article. A PAR of zero indicates good
alignment; higher scores (rarely beyond 50) indicate the
level of irregularity. In this study, the components of the
PAR were recategorized by joining the left and right
buccal occlusion into one subcomponent, named lateral
occlusion. The overjet component was split up into two
subcomponents, overjet and anterior crossbite. The
overbite component was split up into two subcomponents, the overbite and the open bite, thus resulting in
the following eight subcomponents: upper contact point
displacement, lower contact point displacement, lateral
occlusion, overjet, anterior crossbite, overbite, open

302 Al Yami, Kuijpers-Jagtman, and Van t Hof

American Journal of Orthodontics and Dentofacial Orthopedics

March 1999

Fig 1. A and B, Mean nonweighted PAR subscores at the pretreatment stage (TP), posttreatment (T00), postretention
(T0), and the years postretention (T2-T10).

bite, and centerline. All subcomponents were weighted

according to the British weighting factors.18
Three examiners were incorporated in this study. To
determine the measurement error in the PAR and to
assess the intraobserver and interobserver agreement, 18
randomly selected patients were evaluated by the three
observers. The dental casts at TP and at T5 were remeasured for these patients. The time interval between two
intraobserver assessments was at least 3 months.
Statistics

Systematic differences between observers were

tested by the paired t test. Interobserver and intraobserver reliability was expressed as Pearsons correlation coefficients between remeasurements. The magnitude of the intraobservers and interobservers measurement error in the PAR was calculated.
The drop-out analysis included the chi-square test to
compare drop-outs with the remaining patients for sex and
Angle classification, and the t test to compare drop-outs
with the remaining patients for PAR at TP, T00, and T0.

For all stages, the mean and SDs of the PAR were
calculated as well as the absolute and percentage
changes. The mean percentage change per year
(relapse) related to T0 was calculated. Multiple regression analysis was performed to test the effect of the use
of a fixed retainer corrected for confounders.
RESULTS
Reproducibility

No significant systematic differences were found

between examiners (paired t test). The measurement
errors were 0.8 and 1.9 PAR points for the intraobserver and the interobserver measurements, respectively.
The intraobserver correlation ranged over the two periods from 0.98 to 0.99 and the interobserver correlation
from 0.92 to 0.99, indicating a high level of reliability.
Drop-out Analysis

No significant differences were found between the

drop-out and remaining patients for sex and PAR at TP,
T00, and T0. The chi-square test showed a significant-

American Journal of Orthodontics and Dentofacial Orthopedics

Volume 115, Number 3

ly higher drop-out (P = .005) for Class II Division 2

cases compared with other Angle classes.
PAR and Follow-up Stages

Table II shows the mean and SDs for the PAR at the
pretreatment stage and the means and SDs as well as
the absolute and percentage change with respect to TP.
The mean PAR at the pretreatment stage was 28.4
10.2 and at the posttreatment stage 8.5 6.7. During
the postretention period of 10 years, the mean PAR
score increased gradually to 14.6 9.7. The mean percentage change compared with TP was 67.1 27.1 at
the posttreatment stage and 45.2 36.3 at 10 years
postretention. The highest posttreatment mean percentage change per year (relapse) was found during the first
2 years postretention and was 24% per year of the total
relapse. From T5 on, a marginal change was observed
(Table III).
PAR Subscores

Fig 1 shows the mean values for the different PAR

subscores per stage. It shows an immediate decrease in
all PAR subscores because of treatment except for the
anterior open bite where the mean posttreatment score
was higher than the pretreatment score. The score for
the lower anterior contact point displacement shows the
largest changes after treatment. The score was deteriorating up to 10 years postretention. The score at T10
(2.5 2.8) was even higher than at the start of treatment (2.0 2.9).
PAR and Retention Type

Preanalysis showed that the initial (TP) subscores

upper and lower contact point displacement and
open bite as well as upper contact point displacement and open bite at T00 are correlated with the
choice for fixed retention and are potential confounders by indication. Analysis of covariance (correcting for above variables) showed that the presence of
fixed retention had a significant influence on the relapse
at 5 and 10 years postretention. In cases with fixed
retention, the relapse was 3.6 PAR points less at 5 years
postretention and 4.6 PAR points less at 10 years postretention as compared to cases without fixed retention.
DISCUSSION

This study was designed as a longitudinal retrospective study (historical cohort study). Generally
speaking, it is very difficult to avoid selectivity of dropout in such a design. Indeed, a difference in Angle classification between drop-out patients and remaining
patients was found. More Class II Division 2 patients
were lost to follow-up than patients in the other Angle

Al Yami, Kuijpers-Jagtman, and Van t Hof 303

classes. We have no explanation for this finding. A

common feature of relapse in Class II Division 2 cases
is deepening of the overbite. As more Class II Division
2 cases were lost to follow-up, it might be possible that
the overbite subscore as represented in Fig 1 is an
underestimation of the real relapse of this feature.
Another problem related to longitudinal retrospective
studies might be the presence of a secular trend in the
measured variable. In an earlier study, Al Yami et al23
found that the treatment outcome could be explained to
some extent by the treatment period; more recent periods produced better quality. The difference between
periods, however, was only 1 PAR point, which should
be considered as clinical irrelevant.
In this study, the PAR index was used as a measurement for orthodontic treatment outcome. This
index has certain limitations as it measures only
occlusal changes which, although important, are not
the only factors in orthodontic treatment. Factors like
decalcification, root resorption, gingival recession,
inclination of the incisors, and facial aesthetics
undoubtedly contribute to the quality of treatment.
In this study, the mean reduction in the PAR at the
end of active treatment was 20.3 11.4. Ten years
postretention the reduction was still 13.7 12. This
indicates that 67.5% of the achieved orthodontic treatment results as measured by the PAR index still existed 10 years postretention. On the other hand, only a
45% reduction of the PAR compared with the pretreatment PAR existed. When considering these figures, it
should be taken into account, however, that the time
span is 10 years. This means that the measured PAR at
10 years postretention is not the result of orthodontic
treatment alone but also of physiologic and pathologic
changes in the dentition and surrounding tissues during
those years. It has been shown by Behrents24 and
Schols and Van der Linden25 that considerable craniofacial alteration occurs beyond 17 years of age in
human beings. This is accompanied by compensatory
changes in the dentition. The orthodontist has little
control over these biologic processes.
The achieved results started to relapse even within
the retention period, although this change (0.56 6.4
PAR points) is not clinically relevant. Almost half of
the 10 years relapse occurs in the first 2 years postretention (49%). The mean age at the posttreatment stage
was 15.6 3.0 and at the postretention stage 16.7 3.1.
This indicates that there were cases reaching the
postretention stage while some potential growth was
still present. This remnant of growth may influence the
stability of the result of the orthodontic treatment.10
As shown by the analysis of the subcomponents of
the PAR, all occlusal traits of the original malocclusion

304 Al Yami, Kuijpers-Jagtman, and Van t Hof

showed improvement as a result of treatment except the

anterior open bite, which even increased during treatment. All occlusal traits relapsed gradually over time but
remained stable at later stages with the exception of the
lower anterior contact point displacement, which showed
a fast and continuous increase, even exceeding the initial
score. This phenomenon, described as tertiary crowding,
is well known from other clinical studies.2,26
To prevent anterior lower arch crowding, bonded
retainers are commonly used. In our sample (active
treatment before 1982), 11% of the patients had such a
retainer. The cases with fixed retention show a consistently better alignment at 5 and 10 years postretention,
even while the PAR was higher at the pretreatment
stage. These findings should be interpreted with caution
because of the possibility of confounding by indication at the pretreatment stage. It is common practice
that the decision for retention is partly based on pretreatment characteristics.2,10,26 Comparison of the pretreatment subscores for the cases with and without fixed
retention shows that the highest significant indicator for
considering the case as a fixed retainer case later on is
the severity of the upper and lower anterior crowding as
well as the presence of an anterior open bite.
There is no scientific proof that fixed retainers have
a harmful effect on the hard and soft tissues adjacent to
the wire. Artun27 concluded from his investigation that
long-term use of bonded retainers (1 to 8 years postorthodontic treatment) caused no damage to the teeth
and to the hard and soft tissues adjacent to the wire.
Despite this, the usage of a fixed retainer should be limited to cases with a doubtful prognosis of the orthodontic treatment stability and for those patients who worry
about small changes in the orthodontic treatment results.
CONCLUSION

Sixty-seven percent of the achieved orthodontic

treatment result was maintained 10 years postretention.
About half of the total relapse takes place during the first
2 years after retention. All occlusal traits relapsed gradually over time but remained stable from 5 years postretention on with the exception of the lower anterior contact point displacement that showed a fast and continuous increase, even exceeding the initial score. The results
of this type of studies enable clinicians to inform their
patients before treatment about treatment limitations in
order to give them more realistic expectations.

American Journal of Orthodontics and Dentofacial Orthopedics

March 1999

REFERENCES
1. King EW. Relapse of orthodontic treatment. Angle Orthod 1974;44:300-15.
2. Little RM, Riedel RA, Artun J. An evaluation of changes in mandibular anterior alignment from 10 to 20 years posttreatment. Am J Orthod Dentofacial Orthop
1988;93:423-8.
3. Fidler BC, Artun J, Joondeph DR, Little RM. Long-term stability of Angle Class II,
division 1 malocclusions with successful occlusal results at end of active treatment.
Am J Orthod Dentofacial Orthop 1995;107:276-85.
4. Owman G, Bjerklin K, Kurol J. Mandibular incisor stability after orthodontic treatment in the upper arch. Eur J Orthod 1989;11:341-50.
5. Rossouw PE, Preston CB, Lombard CJ, Truter JW. A longitudinal evaluation of the
anterior border of the dentition. Am J Orthod Dentofacial Orthop 1993;104:146-52.
6. Harris EF, Vaden JL. Post-treatment stability in adult and adolescent orthodontic
patients: a cast analysis. Int J Adult Orthod Orthognath Surg 1994;9:19-29.
7. De La Cruz A, Sampson P, Little RM, Artun J, Shapiro PA. Long-term changes in arch
form after orthodontic treatment and retention. Am J Orthod Dentofacial Orthop
1995;107:518-30.
8. Elms TN, Buschang PH, Alexander RG. Long-term stability of Class II, Division 1,
nonextraction cervical face bow therapy, I, model analysis. Am J Orthod Dentofacial
Orthop 1996;109:271-6.
9. Elms TN, Buschang PH, Alexander RG. Long-term stability of Class II, Division 1,
nonextraction cervical face bow therapy, II, cephalometric analysis. Am J Orthod
Dentofacial Orthop 1996;109:386-92.
10. Nanda RS, Nanda SK. Consideration of dentofacial growth in long-term retention and
stability: is active retention needed? Am J Orthod Dentofacial Orthop 1992;101:
297-302.
11. Ahlgren J. A ten-year evaluation of the quality of orthodontic treatment. Swed Dent J
1993;17:201-9.
12. Ghafari J, Jacobsson-Hunt U, Markowitz DL, Shofer FS, Laster LL. Changes of arch
width in the early treatment of Class II, Division 1 malocclusion. Am J Orthod Dentofacial Orthop 1994;106:496-502.
13. Hansen K, Lemamnueisuk P, Pancherz H. Long-term effects of the Herbst appliance
on the dental arches and arch relationships: a biometric study. Br J Orthod
1995;22:123-34.
14. Otuyemi OD, Jones SP. Long-term evaluation of treated Class II Division 1 malocclusions utilizing the PAR index. Br J Orthod 1995;22:171-8.
15. Brin I, Ben-Bassat Y, Blustein Y, Ehrlich J, Hochman N, Marmary Y, Yaffe A. Skeletal and functional effects of treatment for unilateral posterior crossbite. Am J Orthod
Dentofacial Orthop 1996;109:173-9.
16. Dellinger EL. Active vertical corrector treatment: long term follow-up of anterior open
bite treated by the intrusion of posterior teeth. Am J Orthod Dentofacial Orthop
1996;110:145-54.
17. OBrien KD, Shaw WC, Roberts CT. The use of occlusal indices in assessing the provision of orthodontic treatment by hospital orthodontic services of England and
Wales. Br J Orthod 1993;20:28-38.
18. Richmond S, Shaw WC, OBrien KD, Buchanan I, Jones R, Stephens CD, et al. The
development of the PAR index (Peer Assessment Rating): reliability and validity. Eur
J Orthod 1992a;14:125-39.
19. Richmond S, Shaw WC, Roberts CT, Andrews W. The PAR index (Peer Assessment
Rating): Methods to determine outcome of orthodontic treatments in terms of
improvement and standards. Eur J Orthod 1992b;14:180-7.
20. Buchanan IB, Shaw WC, Richmond S, OBrien KD, Andrews M. A comparison of the
reliability and validity of the PAR index and Summers occlusal index. Eur J Orthod
1993;18:27-31.
21. Fox NA, Chadwick SC. The first 100 cases of orthodontic treatment: one year out of
retention. Dent Update 1994;21:288-97.
22. Kahl-Nieke B, Fischbach H, Schwarze CW. Treatment and post-treatment changes in
dental arch width dimensions: a long-term evaluation of influencing co-factors. Am J
Orthod Dentofacial Orthop 1996;109:368-78.
23. Al Yami EA, Kuijpers-Jagtman AM, Vant Hof MA. Occlusal outcome of orthodontic
treatment. Angle Orthod 1998;68:439-44.
24. Behrents RG. Growth in the aging craniofacial skeleton. Monograph 17. Craniofacial
Growth Series. Ann Arbor: Center for Human Growth and Development; 1985.
25. Schols JGJH, Van der Linden FPGM. Gebissentwicklung und Gesichtswachstum in
der Adoleszenz (Development of the dentition and facial growth during adolescence).
Inform Orthod Kieferorthop 1988;68:439-44.
26. Artun J, Garol JD, Little RM. Long-term stability of mandibular incisors following
successful treatment. Angle Orthod 1996;66:229-38.
27. Artun J. Caries and periodontal reactions associated with long-term use of different
types of bonded lingual retainers. Am J Orthod 1984;86:112-8.