Content uploaded by Kevin D O'Brien
Author content
All content in this area was uploaded by Kevin D O'Brien on Dec 21, 2018
Content may be subject to copyright.
ORIGINAL ARTICLE
Effectiveness of treatment for Class II
malocclusion with the Herbst or Twin-block
appliances: A randomized, controlled trial
Kevin O’Brien, PhD, MSc, BDS, FDS, DOrthRCSEng,
a
Jean Wright, MSc, BSc,
b
Frances Conboy, MA, BA,
b
YeWeng Sanjie, BDS, MSc,
c
Nicky Mandall, PhD, BDS, FDSRCSEng, MOrthRCSEng,
d
Stephen Chadwick,
BDS, FDSRCSEdin, MOrthRCSEng,
e
Ivan Connolly, BDS, FDSRCPSGlasg, FFDRCSIrel, MOrthRCSEng,
f
Paul
Cook, MDSc, BChD, FDSRCPSGlasg, LDS, FDS, DOrth, MOrthRCSEng,
g
David Birnie, BDS, FDSRCSEdin,
FDS, MOrthRCSEng,
h
Mark Hammond, MSc, BDS, FDS, RCPSGlasg, MOrthRCSEng,
i
Nigel Harradine, MB,
BS, BDS, FDSRCSEdin, MOrthRCSEng,
j
David Lewis, BDS, FDS, DOrthRCSEng, FRSH,
k
Cathy McDade,
BDS, FDSRCSEdin, DOrthRCSEng,
l
Laura Mitchell, MDS, BDS, FDSRCPSGlasg, MOrthRCSEng,
DOrthRCSEng,
m
Alison Murray, BDS, MSc, FDSRCPSGlasg, MOrthRCSEng,
n
Julian O’Neill, BDS, MSc,
FFDRCSIrel, MOrthRCSEng,
o
Mike Read, BDS, FDSRCSEdin, DOrthRCSEng,
l
Stephen Robinson, MSc, BDS,
FDSRCPSGlasg, MOrthRCSEng,
h
Dai Roberts-Harry, MSc, BDS, FDSRCPSGlasg, MOrthRCSEng,
g
Jonathan
Sandler, MSc, BDS, FDSRCPSGlasg, MOrthRCSEng,
p
and Ian Shaw, PhD, MScD, BDS, FDS, DOrthRCSEng
q
Manchester, Chester, Portadown, Leeds, Portsmouth, Stourbridge, Bristol, Bolton, Bradford, Derbyshire, Kettering,
Chesterfield, and Sunderland, United Kingdom
The aim of this study was to evaluate the effectiveness of Herbst and Twin-block appliances for established
Class II Division I malocclusion. The study was a multicenter, randomized clinical trial carried out in
orthodontic departments in the United Kingdom. A total of 215 patients (aged 11-14 years) were randomized
to receive treatment with either the Herbst or the Twin-block appliance. Treatment with the Herbst appliance
resulted in a lower failure-to-complete rate for the functional appliance phase of treatment (12.9%) than did
treatment with Twin-block (33.6%). There were no differences in treatment time between appliances, but
significantly more appointments (3) were needed for repair of the Herbst appliance than for the Twin-block.
There were no differences in skeletal and dental changes between the appliances; however, the final occlusal
result and skeletal discrepancy were better for girls than for boys. Because of the high cooperation rates of
patients using it, the Herbst appliance could be the appliance of choice for treating adolescents with Class
II Division 1 malocclusion. The trade-off for use of the Herbst is more appointments for appliance repair. (Am
J Orthod Dentofacial Orthop 2003;124:128-37)
This article reports the results of a randomized
clinical trial that evaluated the effectiveness of
orthodontic treatment with either a Herbst or a
Twin-block functional appliance.
Although the provision of early orthodontic growth
modification treatment for Class II malocclusion has
been investigated with randomized trial methodolo-
gy,
1-3
few controlled clinical trials have investigated
the effects of orthodontic growth modification in early
adolescence. These have been confined to evaluating
the effects of the functional appliance phase of treat-
ment.
4-6
The authors of those studies concluded that
a
Professor and chair, Unit of Orthodontics, University Dental Hospital of
Manchester, United Kingdom.
b
Research associate, University Dental Hospital of Manchester.
c
Visiting student (from Wuhan University, People’s Republic of China),
University Dental Hospital of Manchester.
d
Lecturer in orthodontics, University Dental Hospital of Manchester.
e
Consultant orthodontist, Chester Royal Infirmary, previously Blackburn Royal
Infirmary, Chester, United Kingdom.
f
Consultant orthodontist, Craigavon Area Hospital, Portadown, Northern Ire-
land, United Kindom.
g
Consultant orthodontist, Leeds Dental Institute, United Kindom.
h
Consultant orthodontist, Queen Alexandra Hospital, Portsmouth, United
Kindom.
i
Consultant orthodontist, The Corbett Hospital, Stourbridge, United Kingdom.
j
Consultant orthodontist, Bristol Dental Hospital, United Kindom.
k
Consultant orthodontist, Bolton General Hospital, United Kindom.
l
Consultant orthodontist, University Dental Hospital of Manchester, United
Kingdom.
m
Consultant orthodontist, St Luke’s Hospital, Bradford, United Kingdom.
n
Consultant orthodontist, Derbyshire Royal Infirmary, United Kindom.
o
Consultant orthodontist, Kettering General Hospital, United Kindom.
p
Consultant orthodontist, Chesterfield Royal Hospital, United Kindom.
q
Consultant orthodontist, Sunderland Hospital, United Kindom.
Supported by the Medical Research Council (99410454).
Reprint requests to: Professor Kevin O’Brien, Professor and Chair, Orthodontic
Unit, Department of Dental Medicine and Surgery, Higher Cambridge Street,
Manchester M15 6FH, United Kingdom; e-mail, Kevin.O’Brien@man.ac.uk.
Submitted, August 2002; revised and accepted, December 2002.
Copyright © 2003 by the American Association of Orthodontists.
0889-5406/2003/$30.00 ⫹0
doi:10.1016/S0889-5406(03)00345-7
128
most of the correction of the malocclusion was due to
dentoalveolar change and that there was a small but
statistically significant amount of skeletal change.
One disadvantage of removable functional appli-
ances is that extensive cooperation is needed, and
discontinuation rates can vary between 9% and 15%
with the Twin-block.
5,7
One solution to noncompliance
is to use fixed functional appliances, such as the Herbst
appliance.
8
It has been suggested that treatment that
does not depend on compliance has become more
popular during the last 2 decades.
9
There have, how-
ever, been no randomized trials of the effectiveness of
removable and fixed functional appliances that have
followed the treatment through to completion with
fixed appliance therapy. This was the aim of our study.
This investigation had the null hypothesis that there
is no difference in effectiveness between Twin-block
and Herbst appliances.
MATERIAL AND METHODS
Seventeen hospital-based orthodontic specialists in
the United Kingdom (UK) took part in the study. Each
had undergone basic specialty training followed by 3
years of advanced training in the treatment of severe
malocclusions. All were based in orthodontic depart-
ments working in the National Health Service of the
UK. In this system, the orthodontists are salaried, and
treatment is provided at no direct cost to the patient and
family.
We based our sample size calculation for the
number of patients necessary to achieve 80% power
with an ␣of .05 on a clinically meaningful difference in
peer assessment rating (PAR) scores of 15% between
the study groups.
10
The calculation showed that we
needed to recruit 80 patients into each arm of the study
to account for an estimated noncompletion rate of 15%.
The patient inclusion criteria for this investigation
were overjet ⱖ7 mm, second premolars erupted, and
no craniofacial syndrome.
The protocol was approved by the relevant ethics
committees. We followed the guidelines of the Decla-
ration of Helsinki.
11
When a patient who satisfied the inclusion criteria
attended a study clinic, he or she was invited to enter
the study. When consent was obtained from the child
and the parent, the orthodontist gave patient details to
the study center at Manchester University by telephone.
After initial recording of the data, the patient was
randomized to receive treatment with either a Twin-
block or a Herbst appliance. At the beginning of the
study, random number tables were used to prepare
randomization lists, stratified by center and sex into
permuted blocks.
We used a modification of the original Twin-block
design, shown in Figure 1.
6,12
This appliance consisted
of maxillary and mandibular removable appliances
retained with Adams clasps on the first permanent
molars and first premolars. For additional retention, we
used 0.9-mm ball clasps in the mandibular incisor
interproximal areas and a 0.7-mm maxillary labial bow,
which was only activated when the maxillary incisors
were proclined. The jaw registration was taken with
approximately 7 to 8 mm protrusion and the blocks 7
mm apart in the buccal segments. The steeply inclined
planes interlocked at about 70°to the occlusal plane.
When necessary, compensatory lateral expansion of the
maxillary arch was achieved by means of an expansion
screw that was turned once per week. Reactivation of
the blocks was carried out when necessary. All patients
were instructed to wear the appliance for 24 hours per
day (except during contact or water sports). They were
asked to wear the appliance while eating, if possible.
The patients visited the orthodontic departments
every 4 to 8 weeks. When the overjet was fully reduced,
the operator and patient decided on whether to have a
second phase of fixed appliance therapy. If patients did
not have this second phase, their treatment was finished
by grinding the blocks and reducing the wear of the
Twin-block to permit the occlusion to settle to a good
interdigitation.
12
If they proceeded to a second phase of
treatment, fixed appliances were fitted and the treat-
ment continued until the orthodontist and patient were
satisfied with the final occlusion.
The Herbst appliance used was a cast cobalt chro-
mium design, as described by Pancherz (Fig 2).
8
In this
design, the Herbst framework was extended from the
canines posteriorly to include all the erupted teeth.
Where possible, the occlusion was advanced to an
edge-to-edge relationship. The appliances were ce-
mented with glass ionomer cement. After fitting of the
appliance, preadjusted edgewise fixed appliances were
placed as soon as practicable. The patients were seen
every 4 to 8 weeks. During this phase of treatment, if
the operator thought that it was necessary, the appliance
was advanced with collars placed on the pistons. When
the overjet was fully reduced, the Herbst appliance was
removed and the treatment completed. The fixed appli-
ances were removed when the orthodontist and patient
were satisfied with the final occlusion.
A patient was classified as noncompliant for both
the treatment groups if overjet was not reduced by at
least 10% after 6 months or if he or she broke or
damaged the appliance so that treatment was not
practicable.
Data were collected on the patients at the following
points: data collection 1 (DC1) was completed when
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 124, Number 2O’Brien et al 129
each patient entered the study, and the final data
collection (DC2) occurred when the treatment was
completed. The following were collected by each orth-
odontist and sent to the study coordinating center:
●Study models
●Cephalometric radiographs
●The patient’s postal code, used to obtain data on the
patient’s level of social deprivation, according to the
Carstairs index
13
●A questionnaire that gathered information on the
patient’s perception of the appliance, directed at the
effect of the appliance on (1) speaking, eating,
drinking, and appearance; (2) schoolwork; (3) rela-
tionships with friends; (4) relationships with their
families; and (5) hobbies and interests. This was
given to the patients 4 months after the Herbst or
Twin-block appliances were fitted.
●The number of visits required to complete treatment,
additional appointments for appliance repairs, the
number of appliance repairs made, duration of treat-
ment, and date of birth, obtained from each patient’s
chart
The cephalograms were corrected for magnification
and analyzed with the Pancherz analysis.
8
The study
casts were scored with the PAR with the UK weight-
ings.
10
Cephalograms and study casts were both scored
with the examiner unaware of the group to which the
patient had been allocated. The examiner rescored 30
sets of study casts and 20 cephalograms, and error was
evaluated with the intraclass correlation coefficient
(ICC) and Student ttest. This showed no bias for the
PAR index and 0.92 for the ICC. The ICC for cepha-
lometric landmark identification and digitizing ranged
from 0.89 for position of the mandibular base (Pg/OLp)
to 0.97 for position of maxillary central incisor (Is/
OLp) and position of mandibular central incisor (Ii/
OLp). The root mean square (standard deviation of the
error) ranged from 0.51 mm for position of the maxil-
lary base (A/OLp) to 0.81 for Pg/OLp. These were
acceptable levels of error.
We also recorded the stages of maturation of the
cervical spine from the pretreatment cephalograms,
according to the method described by Hassel and
Farman.
14
Thirty sets of radiographs were reanalyzed,
and error was evaluated with the statistic, giving a
value of 0.94; this was acceptable.
Data analysis was performed with SPSS 10.0
(SPSS, Chicago, Ill) and was restricted to generation of
descriptives and regression analyses on (1) the process
of treatment; (2) factors influencing whether the patient
completed the functional appliance phase of the treat-
ment; (3) the final anteroposterior skeletal discrepancy,
as calculated by the Pancherz analysis (defined as
A/OLp ⫺Pg/OLp); (4) the posttreatment overjet; and
(5) the final PAR score.
We carried out an intention-to-treat analysis of
the data, and the results of all patients were analyzed
Fig 1. Design of Twin-block used in study.
Fig 2. Design of Herbst appliance used in study.
American Journal of Orthodontics and Dentofacial Orthopedics
August 2003
130 O’Brien et al
regardless of the outcome of treatment. Details of the
type of regression and the independent variables
assessed during the modeling process are shown in
Table I. No interim modeling of the data was carried
out.We initially considered center ⫻treatment group
and gender ⫻treatment group interaction terms in all
models.
15
Simpler models were then found by re-
moving nonsignificant variables. When variables
were removed, the regression coefficients were com-
pared with the previous model to ensure stability of
effect.
Data on the patients’perceptions of their appliances
were analyzed with the Wilcoxon-Mann-Whitney test.
RESULTS
A total of 215 patients were enrolled in the study: 110
(62 girls and 48 boys) were allocated to receive treatment
with the Twin-block, and 105 (55 girls and 50 boys) to the
Herbst group (Fig 3). Enrollment started in March 1997
and was completed by June 1998. DC2 was done in
September 2001. The average age of the children was
12.41 (95% confidence interval [CI] 12.17-12.63) and
12.74 (95% CI 12.48-12.99) years for the Twin-block and
Herbst appliance groups, respectively. The number of
patients entered by each department ranged from 4 to 39.
In the largest department, 2 operators treated the patients.
The mean deprivation scores for the patients ranged from
⫺1.18 to 3.68, with high scores representing higher levels
of deprivation.
Details of the treatment process are given in Table II.
Analysis with the Wilcoxon-Mann-Whitney test showed
no difference in the total duration of treatment between
the Herbst and the Twin-block groups (P⫽.53). How-
ever, the patients who wore the Twin-block appliance
spent more time in the functional appliance phase of
treatment (P⬍.0005), and they had slightly more routine
appointments (P⬍.04). When we considered the number
of additional appointments that were needed because of
breakage or debonding of the appliances, it seemed that
there were considerably more appointments for the Herbst
appliance group (P⬍.0005).
The regression for log duration of the time of the
functional appliance phase is shown in Table III. We
found a significant interaction between center and
duration of the functional phase. This was studied by
dividing the centers into 2 groups according to the
duration of the functional phase. This led to the
following findings:
●The use of a Twin-block appliance increased the
duration of treatment by a factor of 2.2 months
compared with the Herbst appliance in centers with
shorter treatment times, compared with a factor of
1.5 months in centers with longer duration.
●Patients who were classified as being at cervical
spine development stage CVs1 spent 1.4 times longer
in their functional appliance than those in the later
developmental stages.
●There was a correlation between socioeconomic
status and duration of treatment regardless of
appliance.
Table I. Details of variables that were entered into different regression analyses
Independent variables assessed at start of
modeling process
Type of regression
Logistic
regression Linear regression (Sums of squares Type II)
Failure to
complete
(yes/no)
Natural logarithm of
duration of functional
treatment in months
Skeletal discrepancy
measured by
Pancherz analysis Overjet PAR score
Baseline value VV
Treatment group F(2) F(2) F(2) F(2) F(2)
Center F(13) F(13) F(13) F(13) F(13)
Gender F(2) F(2) F(2) F(2) F(2)
Age F(3) V
Carstairs social deprivation index F(2) V V V V
Pretreatment cephalometric values
(A/OLp, Pg/OLp, max/mand plane) VVV
Time from registration to DC2
cephalograms VVV
Spine maturation F(4) F(4) F(4) F(4) F(4)
Treatment ⫻center F(2) ⫻F(13) F(2) ⫻F(13) F(2) ⫻F(13) F(2) ⫻F(13) F(2) ⫻F(13)
Gender ⫻center F(2) ⫻F(13) F(2) ⫻F(13) F(2) ⫻F(13) F(2) ⫻F(13) F(2) ⫻F(13)
F, Factor (number of levels); V, continuous variable.
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 124, Number 2O’Brien et al 131
The 2 appliances differed significantly in cost. The
average costs were $350 for the Herbst and $80 for the
Twin-block.
We found that 37 (33.6%) of the children in the
Twin-block group and 18 (12.9%) of the Herbst pa-
tients did not complete the functional appliance phase
of treatment (P⫽.01). None of the noncompliant
patients received a second phase of fixed appliance
treatment. The regression analysis of this data is shown
in Table IV.
This shows that fitting with a Twin-block appliance
increased a patient’s chance of not completing the
functional appliance phase of treatment by 2.4 times,
compared with a Herbst appliance. There was also an
effect of the patient’s level of social deprivation. This
suggested that a child in the least-deprived quartile of
our population had 4 times the chance of completing
treatment than did a child in the most deprived quartile,
regardless of the appliance.
By using a cutoff value of the predicted probability
from the regression of 0.35, 58.7% of those not com-
pleting and 70.1% of those completing functional
treatment were correctly predicted by this regression.
Data suggest that the patients treated with the
Twin-block thought that their speech and sleep patterns
had changed, and they felt embarrassed about their
Fig 3. Flow chart of patients in study.
American Journal of Orthodontics and Dentofacial Orthopedics
August 2003
132 O’Brien et al
appliance. Importantly, they also believed that these
factors influenced their relationships with family
(Mann-Whitney P⬍.001). When we evaluated these
factors for the Twin-block patients who did not com-
plete the functional appliance phase of treatment, it
seemed that they felt that the problems with eating
influenced their schoolwork, and they were more em-
barrassed with their families than were the patients who
completed the functional phase (Mann-Whitney P⬍
.001). Similarly, Herbst appliance patients who did not
complete phase I treatment reported more problems
with eating (Mann-Whitney P⬍.005).
Analysis of the cervical spine maturational data
showed that the patients could be classified into the
following groups: Cvs1 ⫽19 (8%), Cvs2 ⫽53 (24%),
Cvs3 ⫽66 (30%), Cvs4 ⫽40 (18%), Cvs5 ⫽1
Table II. Treatment process data
Twin-block Herbst P(Mann-Whitney)
Number of visits
Total 16.05 (14.35 to 17.76) 20.21 (18.57 to 21.86) .0005
n⫽56 n ⫽70
Functional phase 8.64 (7.71 to 9.57) 9.09 (8.24 to 9.93) .0005
n⫽50 n ⫽70
Regular visits 5.63 (5.00 to 6.25) 4.50 (4.05 to 4.94) .04
Emergency visits 1.55 (1.02 to 2.09) 4.29 (3.51 to 5.06) .0005
Fixed phase—only patients receiving fixed phase 11.53 (9.95 to 13.10) 12.77 (11.48 to 14.06) .23
n⫽36 n ⫽61
Time in treatment (months)
Functional phase 11.22 (9.58 to 12.86) 5.81 (5.13 to 6.48) .0005
n⫽56 n ⫽70
Fixed phase—only patients receiving fixed phase 14.81 (12.63 to 16.99) 16.29 (14.57 to 18.01) .292
Total treatment time 21.99 (19.50 to 24.49) 20.84 (18.88 to 23.27) .53
Data are presented as mean (95% CI).
Table III. Results of regression duration of treatment when functional appliance was in place (months) as
dependent variable
Significant variables Coefficient 95% CI for coefficient P
Treatment group (Twin-block) 2.17 1.71 to 2.75 ⬍.0005
Center (1) 1.43 1.11 to 1.83 .044
Cervical spine staging (1) 1.42 1.03 to 1.95 .029
Carstairs social deprivation index 0.97 1.57 to 0.99 .038
Treatment (Twin-block) ⫻center (1) 0.68 0.46 to 1.00 .049
Constant 3.96 3.29 to 4.76 ⬍.005
Time in functional phase (months) Twin-block (n ⫽56) Herbst (n ⫽70)
Center ⫽1 12.01 (8.75 to 15.27) 7.09 (5.93 to 8.25)
Center ⫽2 10.70 (8.88 to 12.53) 4.90 (4.17 to 5.63)
Log duration of treatment was used in the model. These figures have been converted back from log scale.
Table IV. Regression analysis on dependent variables of patients who did not complete functional appliance phase
of treatment
Significant variables Regression
coefficient
Odds ratio for
noncompletion of
functional treatment 95% CI of odds ratio P
Treatment (Twin-block) 0.868 2.38 1.178 to 4.83 .018
Deprivation score .087
Level 1 (lowest) ⫺1.376 0.253 0.08 to 0.77
Level 2 ⫺0.150 0.861 0.35 to 2.12
Level 3 ⫺0.089 0.915 0.38 to 2.23
Compared with Level 4 (highest deprivation score) 0 1
Constant ⫺1.315 0.268 .001
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 124, Number 2O’Brien et al 133
(0.004%), and Cvs 6 ⫽0. It was not possible to
categorize 35 patients (16%) because the radiographs
were not clear in the cervical spine area. The cephalo-
metric data at the start and end of treatment are shown
in Table V, and Table VI contains the data for cepha-
lometric change.
The mean pretreatment PAR scores were 34 (95%
CI 31.74-36.25) for the Twin-block group and 31.14
(95% CI 28.92-33.36) for the Herbst group. At the end
of all treatment, the mean scores were 10.57 (95% CI
7.86-13.28) for the Twin-block group and 7.28 (95% CI
5.87-8.70) for the Herbst patients. When we considered
Table V. Pancherz analysis variables at start and end of study
Twin-block (n ⫽63) Herbst (n ⫽67)
Before After Before After
Mean 95% CI Mean 95% CI Mean 95% CI Mean 95% CI
Overjet
(Is/OLp ⫺Ii/OLp) ⫹10.29 ⫹9.77 to ⫹10.81 ⫹4.05 ⫹3.42 to ⫹4.68 ⫹9.33 ⫹8.76 to ⫹9.90 ⫹3.53 ⫹3.20 to ⫹3.85
Molar relation (Ms/
OLp ⫺Mi/OLp) ⫹1.96 ⫹1.52 to ⫹2.42 ⫺1.66 ⫺2.29 to ⫺1.02 ⫹1.27 ⫹0.77 to ⫹1.76 ⫺1.76 ⫺2.35 to ⫺1.17
Maxillary base
(A point to OLp) ⫹71.46 ⫹70.50 to ⫹72.42 ⫹73.31 ⫹72.11 to 74.51 ⫹72.10 ⫹70.92 to ⫹73.19 ⫹73.33 ⫹72.06 to ⫹74.60
Mandibular base
(Pg/OLp) ⫹71.72 ⫹70.49 to ⫹72.95 ⫹76.14 ⫹74.65 to ⫹77.64 ⫹72.56 ⫹71.09 to ⫹73.83 ⫹76.22 ⫹74.55 to ⫹77.88
Skeletal discrepancy
(A point to OLp ⫺
Pg/Olp)
⫺0.26 ⫺1.03 to ⫹5.09 ⫺2.83 ⫺3.68 to ⫹1.99 ⫺0.46 ⫺1.32 to ⫹0.41 ⫺2.89 ⫺3.68 to ⫺1.92
Condylar head
(Co/OLp) ⫺13.40 ⫺14.08 to ⫺12.71 ⫺14.36 ⫺15.14 to ⫺13.57 ⫺13.22 ⫺13.93 to ⫺12.55 ⫺13.52 ⫺14.23 to ⫺12.80
Mandibular length
(Pg/OLp ⫹
Co/OLp)
⫹58.32 ⫹56.90 to ⫹59.74 ⫹61.78 ⫹60.01 to ⫹63.56 ⫹59.34 ⫹57.64 to ⫹60.79 ⫹62.70 ⫹60.83 to ⫹64.56
Maxillary incisor (Is/
OLp ⫺Ss/OLp) ⫹9.40 ⫹8.93 to ⫹9.87 ⫹6.29 ⫹5.6 to ⫹7.02 ⫹8.50 ⫹8.00 to ⫹8.99 ⫹6.07 ⫹5.43 to ⫹6.72
Mandibular incisor
(Ii/OLp ⫺Pg/OLp) ⫺1.15 ⫺1.96 to ⫺0.37 ⫺0.59 ⫺1.54 to ⫹0.35 ⫺1.29 ⫺2.19 to ⫺0.38 ⫺0.35 ⫺1.27 to ⫹0.57
Maxillary molar
(Ms/OLp ⫺
Ss/OLp)
⫺21.48 ⫺22.07 to ⫺20.89 ⫺21.08 ⫺21.73 to ⫺20.42 ⫺21.80 ⫺22.30 to ⫺21.30 ⫺21.32 ⫺22.10 to ⫺20.26
Mandibular molar
(Mi/OLp ⫺
Pg/OLp)
⫺23.71 ⫺24.55 to ⫺22.88 ⫺22.25 ⫺23.24 to ⫺21.26 ⫺23.53 ⫺24.30 to ⫺22.76 ⫺22.46 ⫺23.39 to ⫺21.53
Table VI. Change in Pancherz analysis variables
Twin-block Herbst
Mean 95% CI Mean 95% CI
Overjet (Is/OLp ⫺Ii/OLp) ⫺6.24 ⫺5.47 to ⫺7.00 ⫺5.80 ⫺6.42 to ⫺5.18
Molar relation (Ms/OLp ⫺Mi/OLp) ⫺3.62 ⫹5.07 to ⫹6.69 ⫺3.03 ⫺3.53 to ⫺2.53
Skeletal changes
Maxillary base (A point to OLp) ⫹1.85 ⫹1.2 to ⫹2.49 ⫹1.22 ⫹0.70 to ⫹1.74
Mandibular base (Pg/OLp) ⫹4.42 ⫹3.63 to ⫹5.20 ⫹3.66 ⫹2.89 to ⫹4.43
Condylar head (Co/OLp) ⫺0.96 ⫺0.51 to ⫺1.39 ⫺0.30 ⫺0.67 to ⫹0.08
Mandibular length (Pg/OLp ⫹Co/OLp) ⫹3.46 ⫹2.45 to ⫹4.47 ⫹3.36 ⫹2.51 to ⫹4.21
Dental changes
Maxillary incisor (Is/OLp ⫺Ss/OLp) ⫺3.11 ⫺2.31 to ⫺3.91 ⫺2.43 ⫺3.00 to ⫺1.85
Mandibular incisor (Ii/OLp ⫺Pg/OLp) ⫹0.56 ⫺0.29 to ⫹1.13 ⫹0.94 ⫹0.37 to ⫹1.51
Maxillary molar (Ms/Olp ⫺Ss/OLp) ⫹0.40 ⫺0.21 to ⫹1.02 ⫹0.48 ⫺0.19 to ⫹1.14
Mandibular molar (Mi/OLp ⫺Pg/OLp) ⫹1.45 ⫹0.66 to ⫹2.55 ⫹1.07 ⫹0.61 to ⫹1.53
American Journal of Orthodontics and Dentofacial Orthopedics
August 2003
134 O’Brien et al
change in score, the Twin-block group had a mean
percentage reduction of 40 (SD ⫽29.3) and the Herbst
patients had a decrease of 39 (SD ⫽21.1).
None of the variables that we considered for the
final overjet regression was significant. However, a
satisfactory model was formulated for the final skeletal
discrepancy; this is shown in Table VII. This shows
that the final skeletal discrepancy was influenced by the
pretreatment discrepancy. It seems that treatment con-
tributes to reducing the discrepancy, but it did not
totally eliminate it. Importantly, the sex of the patients
had an effect in the model, suggesting that girls had less
skeletal II discrepancy than did boys at the end of
treatment, regardless of the appliances.
We also constructed a satisfactory model for final
PAR score (R
2
⫽0.35, P⫽.001). The only variable
that had an effect in this model was gender, with a 
coefficient of 6.5 (95% CI 2.64-10.43).
DISCUSSION
The likelihood of patient cooperation is one of the
most important factors influencing the choice of orth-
odontic treatment. This randomized clinical trial
showed that cooperation with the Herbst appliance was
greater than that with the Twin-block. The noncomple-
tion rate with the Twin-block was twice that of the
Herbst.
A considerable advantage of a prospective study is
that the dropout or failure rate can be accurately
measured. Other prospective investigations show drop-
out rates with Twin-block appliances of 15%,
5
17%,
7
and even 50%.
16
Although the level of noncompliance with the
Twin-block was disappointing, even when the func-
tional appliance was attached to the patient’s teeth, the
discontinuation rate was still rather high. It was not
possible for us to compare our results with other studies
in which compliance depended on treatment because,
surprisingly, noncompliance data have not been report-
ed.
8,9
In our study, the main reason for discontinuation
of treatment was persistent debonding of the Herbst
appliance; this does not seem to be related to any
operator or patient factors. We can therefore conclude
that, even when a functional appliance is fixed to the
teeth, the probability that this treatment will be unsuc-
cessful is 12.9%.
It is difficult to explain the high discontinuation
rates that we found. The most plausible reason must be
that this study was carried out in a “real world”setting,
rather than a dental school with 1 or 2 operators. The
setting of treatment might have had an influence. All
treatment was provided at no cost to the child and
parents; it could be suggested that paying a fee would
ensure cooperation, but this is conjecture.
Another finding of note was the effect of social
deprivation on completion rates of phase I treatment. It
seemed that if a child resides in an area of high social
deprivation, this markedly increased the likelihood of
not completing treatment. This has not been detected in
any other orthodontic investigation, because studies of
factors that influence cooperation have concentrated on
orthodontist-patient interactions.
17,18
Our only other
source of comparable data is to consider the influence
of social deprivation on the uptake of dental care in the
UK. It has been suggested that social deprivation can
result in poor attendance, problems in keeping appoint-
ments, and relapse in oral care; these factors might then
make the operator more likely to stop treatment.
19
Because these findings can probably be extrapolated to
orthodontic treatment, they might explain our findings.
This finding can be considered to be important,
because it suggests that if a study includes only children
from higher socioeconomic backgrounds, the results
are likely to be biased toward successful treatment.
This implies that all studies would have more generality
if an assessment of socioeconomic status were included
as an independent variable.
In all clinical investigations, it is important to
gather information on the perceptions of the consumers
of care. When the patients had worn their appliances for
4 months, we found differences between the 2 appli-
ances. It appeared that the Twin-block, perhaps because
of the bulky acrylic blocks, caused more problems than
did the Herbst in eating and speaking. Arguably, the
patient might repeatedly remove the appliance, thus
influencing the ultimate success of treatment. Further-
more, it seems that the information patients currently
receive (that a new appliance will be uncomfortable for
only a few days) does not reflect their actual experi-
ence.
It seems that the Herbst appliance was more effec-
tive at reducing overjet in phase I of treatment. Unfor-
tunately, this did not lead to shorter overall treatment
times, because the second phase of fixed appliance
treatment was longer. There may be several reasons for
Table VII. Regression analysis on final skeletal
discrepancy according to Pancherz analysis (A/OLp
⫺Pg/OLp)
Significant variables Coefficient 95% CI
for coefficient P
Pretreatment skeletal
discrepancy 0.86 0.72 to 1.0 ⬍.0005
Sex (female) ⫺1.6 ⫺2.5 to ⫺0.7 ⬍.001
For regression analysis, n ⫽147, P⬍.0005, adjusted R
2
⫽0.58.
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 124, Number 2O’Brien et al 135
this; the most likely is that any trimming of the
Twin-blocks during the later months of phase I treat-
ment results in correction of dental features, thus
reducing the complexity of any fixed appliance therapy.
In contrast with the Herbst, the second phase may be
more complex because of persistence of such features
as posterior lateral open bites.
The duration of the functional appliance phase of
treatment was influenced by a factor that was attribut-
able to the treatment center. We could not identify any
variable that could explain this finding, and it appears
that this is an unexplained effect of treatment center.
This is not unusual for multicenter studies and reflects
the pragmatic nature of this investigation.
We were disappointed to find that the Herbst
appliance was prone to damage resulting in debonding
and fractured components. This has not been reported
in the literature. However, an unpublished thesis
showed much higher breakage and debonding rates.
20
As a result, it could be suggested that the trade-off for
the increased compliance rate is that the patient must
return to the clinician for several appliance repairs
during the functional phase of treatment.
Evaluation of the morphologic effects of the appli-
ances shows that most of the changes were dental; the
maxillary incisors were retracted and the mandibular
incisors were proclined. In addition, the skeletal
changes were less than those reported in retrospective
investigations.
8,9
When we consider the dental effects of the appli-
ances, treatment group was not significant in explaining
the final overjet, and we conclude that both treatments
are equally effective at reducing overjet. We did,
however, succeed in fitting a model for the final PAR
score. It appears that there was more residual maloc-
clusion for boys than for girls, and the difference of 6
PAR points is clinically significant. This could have
resulted in differences in cooperation between girls and
boys. This is worth further investigation.
When the data on final skeletal discrepancy are
evaluated, a minus value for the Pancherz skeletal
discrepancy means that a patient is less Class II than if
he or she had a positive value. Interpretation of the
regression for the effect of sex with the coefficients
suggests that, after treatment, girls were 1.6 mm less
Class II than were boys, regardless of whether they
received Twin-block or Herbst treatment.
It is difficult to explain these effects. They could be
due to different developmental stages of the boys and
girls in the study. However, the cervical spine growth
staging did not have an influence in the model. Another
reason could be potential differing levels of cooperation
(as suggested for the difference in PAR scores); nev-
ertheless, this was not counteracted by the use of the
Herbst appliance that did not depend on compliance.
We also found that the initial skeletal discrepancy
influenced the outcome. It appears that the initial
discrepancy was not totally counteracted by the
effects of the treatment. This is a similar finding to
that from our study on the effects of early orthodon-
tic treatment with the Twin-block.
21
In addition, we
also found that when we included maxillary/mandib-
ular plane angle as a measure of vertical proportion
in the regression, it did not have an effect in the
model. This suggests that the orthodontic clinical
perception that patients with reduced facial heights
or large skeletal discrepancies respond better to
functional appliances is not correct.
Another important finding was the possible effects
of stage of maturation of the cervical spine. It has been
suggested that the best time to provide treatment is
between CVs3 and CVs4, because this coincides with
peak growth of the mandible.
22,23
Although the results
of the analysis of the duration of time in functional
appliance data seem to reinforce this finding, this was
not found for our data on skeletal discrepancy, and
neither did this explain the different final skeletal
relationships between girls and boys. As a result, our
data do not support this theory, and it requires investi-
gation in other prospective studies.
The Herbst appliance has some advantages over the
Twin-block, mostly concerning increased compliance.
Nevertheless, the trade off for these benefits is the
additional cost of appliance construction and the extra
visits for appliance repair. This will ultimately deter-
mine the uptake of this technique. This study adds to
our knowledge of the effects of functional appliances;
however, its findings might not be totally applicable to
other countries or health care systems. Other research-
ers should repeat this methodology in other locations.
CONCLUSIONS
From this study, we can conclude that:
1. Patient cooperation with the Herbst appliance is
better than that with the Twin-block.
2. Phase I treatment is more rapid with the Herbst
appliance, but overall duration of treatment is sim-
ilar to that with the Twin-block.
3. The Herbst appliance is prone to debonding and
component breakage.
4. There are no differences in the dental and skeletal
effects of treatment between the 2 appliances, but
there was a marked sex effect: girls responded to
treatment better than boys.
American Journal of Orthodontics and Dentofacial Orthopedics
August 2003
136 O’Brien et al
The authors thank the patients for taking part in this
study and the supporting staff for their additional work
at the treatment centers.
REFERENCES
1. Tulloch JFC, Philips C, Proffit WR. Benefit of early Class II
treatment: progress report of a two-phased randomized clinical
trial. Am J Orthod Dentofacial Orthop 1998;113:62-72.
2. Keeling SD, Wheeler TT, King GJ, Garvan CW, Cohen DA,
Cabassa S, et al. Anteroposterior skeletal and dental changes
after early Class II treatment with bionators and headgear. Am J
Orthod Dentofacial Orthop 1998;113:40-50.
3. Ghafari J, Shofer FS, Jacobsen-Hunt U, Markowitz DL, Laster
LL. Headgear versus functional regulator in the early treatment
of Class II, Division 1 malocclusion. Am J Orthod Dentofacial
Orthop 1998;113:51-61.
4. Nelson CN, Harkness M. Mandibular changes during functional
appliance treatment. Am J Orthod Dentofacial Orthop 1993;104:
153-61.
5. Illing HM, Morris DO, Lee RT. A prospective evaluation of
Bass, bionator and Twin-block appliances. Part 1—the hard
tissues. Eur J Orthod 1998;20:501-16.
6. Lund I, Sandler PJ. The effects of Twin-Blocks: a prospective
controlled study. Am J Orthod Dentofacial Orthop 1998;113:
104-10.
7. Harradine NWT, Gale D. The effects of torque control spurs in
twin-block appliances. Clin Orthod Res 2000;3:202-9.
8. Pancherz H. The mechanism of Class II correction in Herbst
appliance treatment. Am J Orthod 1982;82:104-13.
9. Franchi L, Baccetti T, McNamara JA. Treatment and posttreat-
ment effects of acrylic splint Herbst appliance therapy. Am J
Orthod Dentofacial Orthop 1999;115:429-38.
10. Richmond S, Shaw WC, O’Brien KD, Buchanan R, Jones R,
Stephens CD, et al. The development of the PAR index (Peer
Assessment Rating): reliability and validity. Eur J Orthod 1992;
14:125-39.
11. World Medical Association. Ethical principles for medical re-
search involving human subjects (Declaration of Helsinki).
Available at: http://www.wma.net/e/policy/17-c_e.html. Ac-
cessed May 2, 2003.
12. Clark WJ. Twin-block functional therapy. London: Mosby-
Wolfe; 1995.
13. Carstairs V. Multiple deprivation and health state. Community
Med 1981;3:4-13.
14. Hassel B, Farman AG. Skeletal maturation evaluation using
cervical vertebrae. Am J Orthod Dentofacial Orthop 1995;107:
58-66.
15. Senn S. Some controversies in planning and analysing multi-
centre trials. Stat Med 1998;17:1753-65.
16. Baton S, Cook PA. Predicting functional appliance treatment
outcome in Class II malocclusions—a review. Am J Orthod
Dentofacial Orthop 1997;112:282-6.
17. Sinha PK, Nanda RS, McNeil DW. Perceived orthodontist
behaviors that predict patient satisfaction, orthodontist-patient
relationship, and patient adherence in orthodontic treatment.
Am J Orthod Dentofacial Orthop 1996;110:370-7.
18. Nanda RS, Kieri MJ. Prediction of cooperation in orthodontic
treatment. Am J Orthod Dentofacial Orthop 1992;102:15-21.
19. Pavi E, Kay EJ, Shepherd KW. The effect of social and personal
factors on the utilisation of dental services in Glasgow, Scotland.
Community Dent Health 1995;12:208-15.
20. Tse ELK. A comparative study of the skeletal and dental effects
of a modified Herbst appliance and traditional Herbst in Southern
Chinese—a prospective study [thesis]. Hong Kong: University of
Hong Kong; 1994.
21. O’Brien KD, Wright J, Conboy F, Sanjie Y, Mandall N,
Chadwick S, et al. Effectiveness of early orthodontic treatment
with the Twin-block appliance: a multicenter, randomized, con-
trolled trial: Part I. Dental and skeletal effects. Am J Orthod
Dentofacial Orthop 2003 (in press).
22. Franchi L, Baccetti T, McNamara JA. Mandibular growth as
related to cervical vertebral maturation and body height. Am J
Orthod Dentofacial Orthop 2000;118:335-40.
23. Baccetti T, Franchi L, Toth LR, McNamara JA. Treatment
timing for Twin-block therapy. Am J Orthod Dentofacial Orthop
2000;118:159-70.
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 124, Number 2O’Brien et al 137
