SYSTEMATIC REVIEW

Effect of orthodontic treatment with

4 premolar extractions compared with
nonextraction treatment on the vertical
dimension of the face: A systematic
review
Georgios Kouvelis,a Konstantinos Dritsas,b Ioannis Doulis,c Dimitrios Kloukos,d and Nikolaos Gkantidisb
Athens and Thessaloniki, Greece, and Bern, Switzerland

Introduction: Our aim was to assess the available evidence for the effects of orthodontic treatment with 4 pre-
molar extractions on the skeletal vertical dimension of the face compared with nonextraction treatment.
Methods: Electronic database searches (MEDLINE, EMBASE, Cochrane Oral Health Group's Trials Register,
and CENTRAL) of published and unpublished literature and hand searches of eligible studies were performed,
with no language or publication date restrictions. Two authors performed data extraction independently and in
duplicate. Risk of bias was assessed. Results: After application of the eligibility criteria, 14 studies were included
in this systematic review. All were retrospective. Risk of bias ranged from moderate to critical. Ten studies inves-
tigated patients with various skeletal vertical patterns and classes of malocclusion and found no difference
between extraction (Ex) and nonextraction (Nonex) treatment in regard to the vertical dimension. Only 2 studies
found statistically significant increases in the nonextraction groups, one in N-Me (Ex: 11.5 mm;
Nonex: 15.5 mm; P \0.05) and one in SN-GoGn (Ex:
0.9 ; Nonex: 10.8 ; P \0.05), but without a
concurrent significant change in other vertical measurements such as FMA. Two other studies showed
opposite findings regarding N-Me (Ex: 12.3 mm; Nonex: 10.9 mm; P \0.05) and FMA (Ex: 10.3 ; Nonex:

2.0 ; P\0.05). Conclusions: Although the quality of evidence ranged from moderate to low, there was consid-
erable agreement among these studies, suggesting that orthodontic treatment with 4 premolar extractions has
no specific effect on the skeletal vertical dimension. Thus, an extraction treatment protocol aiming to reduce or
control the vertical dimension does not seem to be an evidence-based clinical approach. (Am J Orthod
Dentofacial Orthop 2018;154:175-87)

E
xtractions for orthodontic purposes have always dimension during orthodontic treatment, this may be an
been a controversial issue in contemporary treatment additional major disputation. It has been advocated that
planning. When linked to the control of the vertical changes in vertical dimension during growth may also in-
a
fluence the anteroposterior position of the mandible and
Department of Orthodontics and Dentofacial Orthopedics, 251 Hellenic Air
Force Hospital, Athens, Greece.
the establishment of the permanent occlusion.1
b
Department of Orthodontics and Dentofacial Orthopedics, University of Bern, For many years, posterior tooth extractions have
Bern, Switzerland.
c
been suggested, especially in long-face patients to con-
Department of Orthodontics, Faculty of Dentistry, Aristotle University of Thessa-
loniki, Thessaloniki, Greece.
trol the vertical dimension.2 It has been recommended
d
Department of Orthodontics and Dentofacial Orthopedics, 251 Hellenic Air that extracting permanent teeth may correct an open
Force Hospital, Athens, Greece; Department of Orthodontics and Dentofacial Or- bite or reduce the vertical dimension of the face by coun-
thopedics, University of Bern, Bern, Switzerland.
The first 2 authors contributed equally to this work.
terclockwise rotation of the mandible, through the for-
All authors have completed and submitted the ICMJE Form for Disclosure of Po- ward movement of the posterior teeth: the wedge-type
tential Conflicts of Interest, and none were reported. effect.3-5 According to various authors, extractions
Address correspondence to: Nikolaos Gkantidis, Department of Orthodontics and
Dentofacial Orthopedics, University of Bern, CH-3010, Freiburgstrasse 7, Bern,
lead to reduction of the vertical dimension not only in
Switzerland; e-mails, nikosgant@yahoo.gr; nikolaos.gkantidis@zmk.unibe.ch. hyperdivergent patients, but also in those with skeletal
Submitted, January 2018; revised and accepted, March 2018. open bite.6,7 On the other hand, other studies reported
0889-5406/$36.00
Ó 2018 by the American Association of Orthodontists. All rights reserved.
no distinct effects of extraction treatment on the facial
https://doi.org/10.1016/j.ajodo.2018.03.007 vertical dimension.8,9
175
176
Hyperdivergent patients comprise a significant part
(22%) of the orthodontic patients treated worldwide.10
Thus, the selection of an extraction vs a nonextraction
treatment protocol in regard to effects in the vertical
dimension is a common decision made in every practice.
The contradictory findings of previous studies, however,
do not allow for evidence-based decision making, lead-
ing, in turn, to extremely different strategies applied to
patients by various clinicians.
Therefore, the aim of this review was to systemati-
cally search the literature to evaluate the effects of 4
premolar extractions compared with nonextraction
treatment on the skeletal vertical dimension of the
face.
MATERIAL AND METHODS
Protocol and registration
The protocol was not registered prior to the study.
Selection criteria applied for the review
Any study design was considered eligible for inclu-
sion in this review, including randomized clinical trials,
nonrandomized or quasi-randomized controlled trials,
and prospective and retrospective studies with a treated
comparison or control group.
Patients of any age having orthodontic treatment
with full fixed appliances in the maxilla and mandible
were eligible.
Orthodontic therapy with fixed appliances including
4 premolar extractions (1 in each quadrant) were eligible.
For the control or comparison group, we chose or-
thodontic therapy with fixed appliances, but without
tooth extractions.
The outcome was the effects on the vertical dimen-
sion of skeletal craniofacial structures, quantitatively as-
sessed on radiographic images.
The follow-ups included before and after orthodon-
tic treatment evaluations.
Exclusion criteria were animal and in-vitro studies,
studies involving orthognathic surgery, and case reports
or studies reporting outcomes from fewer than 10 pa-
tients.
Search strategy for identification of studies
Detailed search strategies were developed and appro-
priately revised for each data base, considering the dif-
ferences in controlled vocabulary and syntax rules. The
following electronic databases were searched: MEDLINE
(via Ovid and PubMed, from 1946 to August 2, 2017;
Appendix), EMBASE (via Ovid), Cochrane Oral Health
Group's Trials Register, and CENTRAL.
August 2018  Vol 154  Issue 2 American Journal of Orthodontics and Dentofacial Orthopedics
Kouvelis et al
Unpublished literature was searched on ClinicalTrials.
gov, the National Research Register, and Pro-Quest
Dissertation Abstracts and Thesis database. We attemp-
ted to identify all relevant studies irrespective of lan-
guage. The reference lists of all eligible studies were
hand searched for additional studies.
Selection of studies
Studies were selected independently and in duplicate
by 2 authors (G.K., K.D.) who were not blinded to the
identity of the authors of the studies, their institutions,
or the results of their research. Study selection proced-
ures comprised reading of titles, abstracts, and full texts.
After they excluded noneligible studies, the full report of
publications considered eligible for inclusion by either
author was obtained and assessed independently. Dis-
agreements were resolved by discussion and consulta-
tions with other authors (I.D., N.G.). A record of all
decisions on study identification was kept.
Data extraction and management
Data extraction was performed independently and in
duplicate by the first 2 authors (G.K., K.D.). To record the
desired information, customized data collection forms
were used. Disagreements were resolved in reevaluations
of the original studies by both authors and discussions
with the last author until consensus was reached.
Measures of treatment effect
For continuous outcomes, mean differences and
standard deviations were used to summarize the data
from each study.
Unit of analysis
In all cases, the unit of analysis was the patient.
Missing data
We contacted the authors via e-mail to request infor-
mation that was missing. In case of no response, only the
available data were reported and analyzed. For missing
standard deviations, data could be retrieved if t statistics
or P values were reported. When P values were reported
as P50.000, we considered them rounded values and
made a worst case scenario hypothesis, imputing the
value to be 0.0004 instead.
Assessment of heterogeneity
Heterogeneity can be divided into 3 forms: clinical,
methodological, and statistical.11 We assessed clinical
heterogeneity by examining the characteristics of the
studies, the similarities between the participants, the
Kouvelis et al
Fig. Flow chart of study selection.
interventions, and the outcomes as specified in the in-
clusion criteria. We determined whether the participants
had completed their growth as participant-specific com-
ponents. We also identified the vertical skeletal pattern
and the sagittal skeletal and dental patterns as potential
population-specific sources of clinical heterogeneity.
The potential effect of varying sex distribution between
studies was ignored, since a differential effect due to sex
was not expected and angular measurements were
mainly considered. On the other hand, treatment proto-
col was considered to add to the total clinical heteroge-
neity, and the following aspects were underlined: type of
anchorage, treatment biomechanics (eg, headgear appli-
cation), type of teeth extracted, and type of control
group. The final component of clinical heterogeneity
was outcome measures, since the vertical changes can
be measured on a lateral cephalogram in various ways.
Methodologic heterogeneity was considered in terms
of both study design and conduct, in the latter case,
American Journal of Orthodontics and Dentofacial Orthopedics
177
especially in terms of risk of bias. Finally, the tool we
used to assess risk of bias provided a summary measure
that also accounted for heterogeneity assessment.12
Assessment of reporting bias
Reporting biases arise when the reporting of research
findings is affected by the nature or direction of the find-
ings themselves. We attempted to minimize potential re-
porting biases, including publication bias, multiple
(duplicate reports) publication bias, and language bias
in this review by conducting an accurate and sensitive
search of many sources with no language restriction.
We also searched for ongoing trials.
Quality assessment
The quality assessments of the included studies were
conducted using the ROBINS-I tool.12 Two authors
(G.K., K.D.) assessed the studies individually and then
August 2018  Vol 154  Issue 2
178
compared their findings. Disagreements were resolved
by discussion with the last author (N.G.).
RESULTS
Description of studies
The flow diagram of study selection is shown in the
Figure. A total of 927 studies were initially identified
in the electronic search. After title and abstract
screening, 34 studies were retrieved to be examined in
more detail. Twenty studies were subsequently excluded
for the reasons stated in the flow chart. Consequently, 14
articles were included in this review.
All included studies were retrospective and investi-
gated the effects of 4 premolar extractions. In 11
studies,8,9,13-21 4 first premolars were extracted; in
another study,22 2 maxillary first and 2 mandibular first
or second premolars were extracted. In 2 studies, the
exact premolar extraction pattern was not specified.23,24
An overview of the main characteristics of the studies is
presented in Table I.
Quality assessment
By definition of the tool that we used, no study
received the grade of “low” with regard to confounding,
because confounding was expected with our topic and
was not fully controlled in any case. Most studies were
deemed problematic due to the lack of information
about the treatment mechanics used to close the extrac-
tion spaces. A judgment between “serious” and “critical”
overall risk of bias was made for studies that did not pro-
vide information for at least 1 bias domain.
Of the 14 retrospective studies, 5 were assessed as
moderate,8,9,13,14,17 5 as serious,15,20-22,24 and 3 as
critical risk of bias.16,18,19 Finally, 1 study was assessed
as serious for the Class I group and critical for the
Class II group.23
Disagreements were all between studies with serious
and critical risks of bias. An overview of the risk of bias
assessment is given in Table II.
Qualitative synthesis of included studies
Clinical heterogeneity was high; for this reason, no
quantitative synthesis was possible. The diversity between
growth stages, skeletal patterns, extraction patterns, con-
trols (Table I), and outcomes (Table III) did not lead to
studies with comparable outcome measures. Most studies
included growing patients, but 3 studies15,18,21 included
mainly nongrowing participants. No study reported the
results separately. Other studies did not report on the
vertical skeletal patterns of their samples. Different
biomechanical strategies were used between studies; in
certain studies, this information was incomplete.
August 2018  Vol 154  Issue 2 American Journal of Orthodontics and Dentofacial Orthopedics
Kouvelis et al
Normodivergent patients
Four studies assessed the effects of 4 premolar
extraction treatment on normodivergent pa-
tients.9,18,20,22 One had moderate,9 2 had serious,20,22
and 1 had critical18 risks of bias. The first study
included Class I and Class II Division 1 patients with 4
first premolar extractions, and no difference was found
in the vertical dimension changes.9 For Class II patients,
1 study investigated the effect of 2 maxillary first and 2
mandibular first or second premolars and also detected
no difference compared with the controls.22 In contrast
to the initial expectations, another study, which tested 4
first premolar extractions, found a statistically signifi-
cant decrease (1.7 ) in the mandibular plane angle of
the nonextraction subjects compared with the extraction
group.20 In Class I patients, 1 study detected a smaller
increase (1.4 mm) in N-Me distance in the nonextraction
group compared with the 4 first premolar extraction
group.18 In both cases, the magnitude of differences
was quite small and clinically questionable.
Hyperdivergent patients
Six studies investigated vertical changes in hyperdi-
vergent patients.8,14,15,17,23,24 Three of them8,14,17
were assessed at moderate and 215,24 at serious risks of
bias. The study of Basciftci and Usumez23 was assessed
as having serious (Class I group) and critical (Class II
Division 1 group) risks of bias.
Three studies examined the effects of extraction
treatment in Class I hyperdivergent patients.14,15,23
Only 1 found a statistically significant increase (4 mm)
in N-Me distance in the nonextraction treatment
compared with the extraction approach.15 However, no
difference was detected in the FMA angle.
Four studies assessed the vertical dimension changes
in Class II hyperdivergent patients.8,17,23,24 One of them
found a statistically significant, although clinically
questionable, decrease (1.7 ) in the SN-GoGn angle in
the 4 first premolar extraction group, compared with
the nonextraction group, whereas no difference was
evident for the FMA angle.17 On the contrary, no differ-
ence was found in the other 3 studies.8,23,24 Moreover,
one of them tested opposite treatment approaches that
were expected to increase the effect in the vertical
dimension.8
Other vertical patterns
One study with a moderate risk of bias included a
balanced sample of Class I, Class II, and Class III female
patients and all vertical patterns; no differences were
found between the 4 first premolar extraction and the
nonextraction groups.13
 American Journal of Orthodontics and Dentofacial Orthopedics

Kouvelis et al
Table I. Main characteristics of the included studies
Extraction group (Ex) Nonextraction group (Nonex)

Author Intervention Intervention Malocclusion/Skeletal Duration of

Year Sample size, age Treatment Sample size, age Treatment vertical pattern treatment Outcome of Magnification
Study design Main objective (mean 6 SD), sex details (mean 6 SD), sex details (pretreatment) (mean 6 SD) interest correction method
Basciftci and Changes in Four premolar NA Nonextraction NA Dental Class I Class I (1) Sn-GoGn NA
Usumez23 craniofacial (NA) extractions treatment and Class II, 1 Ex: 1.9 6 0.3 y ( )
(2003) structures of Class I Class I Hyperdivergent Nonex:
Retrospective Class I and Class n 5 22 (13 F, 9 M) n 5 25 (15 F, Class I 1.6 6 0.4 y
II patients age, 14.5 6 2.0 y 10 M) SN-GoGn: Class II, 1
Class II, 1 age, 13.6 6 1.8 y Ex: 36.9 6 5.8 Ex: 2.0 6 0.5 y
n 5 20 (13 F, 7 M) Class II, 1 Nonex: 35.3 6 6.4 Nonex:
age, 17.4 6 3.6 y n 5 20 (12 F, 8 M) Class II, 1 1.9 6 0.4 y
age, 12.7 6 1.0 y SN-GoGn:
Ex: 39.3 6 5.7
Nonex: 34.7 6 5.4
Bravo et al22 Changes in facial Four first premolar NA Nonextraction NA Dental Class II Ex: 2.7 6 0.3 y (1) FMA ( ) Machine
(1997) profile of Class II or two maxillary treatment Normodivergent Nonex: magnification
Retrospective patients first and two n 5 15 F FMA 2.6 6 0.8 y factor
mandibular age, 12.9 6 0.7 y Ex: 26.5 6 2.5
second premolar Nonex: 25.5 6 3.8
extractions
n 5 16 F
age, 13.5 6 1.8 y
Gkantidis et al8 Changes in vertical Four first premolar “Non-extrusive” Nonextraction “Extrusive” Dental Class II, 1 Ex: 3.2 6 0.6 y (1) SN-GoGn Reference ruler
(2011) dimension of extractions mechanics: treatment mechanics: low- Hyperdivergent Nonex: ( )
Retrospective Class II, 1 n 5 29 (16 F, Nance and n 5 28 (14 F, pull headgear in SN-GoGn: 2.4 6 0.8 y (2) FH-MP ( )
hyperdivergent 13 M) Goshgarian 14 M) all patients, Ex: 38.4 6 3.8 (3) ANS-Me
patients age, 11.8 6 0.9 y arches, no Class age, 11.0 6 1.0 y anterior bite Nonex: 39.1 6 3.8 (mm)
II elastics, no planes, Class II (4) LFH/TFH
low-pull elastics or (ratio)
August 2018  Vol 154  Issue 2

headgear or bite posterior

plates crossbite elastics
when needed
Hayasaki et al9 Changes in Four first premolar Headgear, lip Nonextraction Headgear, lip Dental Class I Class I: (1) SN-GoGn Machine
(2005) craniofacial extractions bumper when treatment bumper when and II, 1 Ex: 2.4 y ( ) magnification
Retrospective structures of Class I: needed Class I: needed Normodivergent Nonex: 2.0 y (2) Me-PP factor
Class I and Class n 5 15 (9 F, 6 M) No skeletal n 5 15 (8 F, 7 M) No skeletal SN-GoGn Class II: (mm)
II patients age, 12.3 y anchorage age, 11.9 y anchorage Class I Ex: 2.9 y (3) N-Me (mm)
Class II: Class II: Ex: 32.3 6 5.1 Nonex: 2.6 y (4) ANS-Me
n 5 14 (7 F, 7 M) n 5 15 (7 F, 8 M) Nonex: 31.4 6 5.3 (mm)
age, 12.3 y age, 12.1 y Class II, 1
Ex: 31.2 6 6.1
Nonex: 31.9 6 3.1

179
August 2018  Vol 154  Issue 2

180
Table I. Continued

Extraction group (Ex) Nonextraction group (Nonex)

Author Intervention Intervention Malocclusion/Skeletal Duration of

Year Sample size, age Treatment Sample size, age Treatment vertical pattern treatment Outcome of Magnification
Study design Main objective (mean 6 SD), sex details (mean 6 SD), sex details (pretreatment) (mean 6 SD) interest correction method
Hosseinzadeh- Changes in the Four first premolar Straight wire Nonextraction Straight wire Class I 24 6 6 months (1) Pog-SN7 Machine
Nik et al21 mandible of extractions appliances treatment appliances Ex: with bimaxillary (mm) magnification
(2016) Class I patients n 5 18 F (0.022 in, MBT) n 5 18 F (0.022 in, MBT) protrusion (2) B-SN7 factor
Retrospective age, 16.4 6 0.4 y age, 16.4 6 0.4 y Nonex: without bimaxillary (mm)
protrusion
Katsaros et al24 Changes in the soft Four premolar NA Nonextraction NA Overjet $ 6 mm Ex: 3.7 (1.1- (1) SN-GoMe NA
(1996) tissue profile of extractions treatment Hyperdivergent 5.9) y ( )
Retrospective patients with n 5 33 F n 5 29 F SN-GoMe Nonex: 3.1 (2) PP-GoMe
enlarged overjet age, 13.0 (12.0- age, 12.1 (12.0- Ex: 38.6 6 6.9 (1.3-5.5) y ( )
15.0) y 16.0) y Nonex: 36.3 6 4.8
Kirschneck Changes in vertical Four first premolar Headgear and Nonextraction Headgear, Class II/ Dental Classes 2.8 (2.3-4.8) y (1) SN-GoGn Reference ruler
et al13 dimension extractions Class II/III treatment III elastics I, II, and III ( )
(2015) n 5 25 F elastics n 5 25 F All patterns (2) PP-GoGn
Retrospective age, 10.8 6 1.2 y age, 11.1 6 1.3 y SN-GoGn ( )
American Journal of Orthodontics and Dentofacial Orthopedics

Ex: 35.1 6 5.1

Nonex: 34.2 6 4.8
Kocadereli14 Changes in vertical Four first premolar No extraoral Nonextraction No extraoral Dental Class I NA (1) SN-GoGn NA
(1999) dimension of extractions appliances treatment appliances Hyperdivergent (tendency) ( ) (same radiographic
Retrospective Class I patients n 5 40 (23 F, Minimal posterior n 5 40 (24 F, SN-GoGn (2) FMA ( ) unit)
17 M) protraction 16 M) Ex: 36.9 6 5.5 (3) N-Me (mm)
age, 12.8 6 2.4 y expected due to age, 12.3 6 2.2 y Nonex: 34.7 6 6.4 (4) ANS-Me
initial crowding (mm)
Kumari and Changes in vertical Four first premolar No functional Nonextraction No functional Skeletal Class I NA (1) FMA ( ) NA
Fida15 dimension of extractions appliances treatment appliances Hyperdivergent (tendency) (2) N-Me (mm)
(2010) Class I patients n 5 40 n 5 41 FMA (3) N-ANS/
Retrospective age, 15.4 6 1.2 y age, 15.8 6 1.5 y Ex: 27.8 6 6.1 ANS-Me
Nonex: 26.3 6 6.0 (ratio)
Luppanapornlap Changes in vertical Four first premolar NA Non-extraction NA Class II NA (1) FMA ( ) NA
and dimension of extractions treatment Normodivergent (Ex) and (2) N-Me (mm)
Johnston16 Class II patients n 5 33 (18 F, n 5 29 (18 F, hypodivergent (Nonex) (3) ANS-Me
(1993) 15 M) 11 M) FMA (mm)
Retrospective age, NA age, NA Ex: 24.7
Nonex: 18.9
Paquette et al20 Changes in vertical Four first premolar NA Nonextraction NA Dental Class II, 1 Ex: 1.8 y (1) FMA ( ) NA

Kouvelis et al
(1992) dimension of extractions treatment Normodivergent Nonex: 1.6 y (2) N-Me (mm)
Retrospective Class II patients n 5 33 (20 F, n 5 30 (11 F, Ex: 22.9 (3) ANS-Me
13 M) 19 M) Nonex: 24.9 (mm)
age, 12.5 y age, 12.6 y
 Kouvelis et al 181

Another study with a critical risk of bias compared

correction method
Magnification

magnification
Class II normodivergent extraction patients with Class
II hypodivergent nonextraction patients and also found
no difference.16

(1) GoMe-FH Machine

(2) N-Me (mm) factor

Two studies with serious and critical risks of bias did
Ex: 2.5 6 0.7 y (1) SN-GoGn NA

(1) MP-HP ( ) NA
not define the vertical growth pattern of their sam-

(2) N-Me (mm)

ples.19,21 One tested the effects of 4 first premolar
(3) N-S-Gn ( )
Outcome of

(3) ANS-Me
interest

2.2 6 0.4 y (2) FMA ( )

(3) N-ANS/
ANS-Me
extractions in Class I patients with bimaxillary

(ratio)
(mm)
protrusion, compared with nonextraction treatment in
( )

( )

patients without protrusion and did not detect any

difference in the vertical position of the Point B or Pog
(mean 6 SD)
Duration of

Nonex: 2.9 y
treatment

landmarks. The second investigated Class I patients

Ex: 1.5 y

and also found no difference between groups.

Nonex:

NA

DISCUSSION
Malocclusion/Skeletal

The control of vertical dimension has always been a

vertical pattern
(pretreatment)

Nonex: 38.0 6 3.1

critical issue in orthodontics. The vertical dimension is

SN-GoGn: 32 6 1
Ex: 40.0 6 3.6
Dental Class II, 1

an important component of the human face and often

Normodivergent
Hyperdivergent

Dental Class I

Dental Class I

plays a major role in treatment planning. For example,

SN-GoGn

in hyperdivergent patients, treating doctors are con-

cerned to decrease or at least not to increase the verti-
cal dimension of the face and thus resort to certain
high-pull headgear
for anchorage in

strategies that are considered intrusive, such as tooth

intermaxillary
Nonextraction group (Nonex)

elastics, and
Treatment

extractions or high-pull headgear. The opposite is

all patients
details
Lip bumper,

true for patients with a short lower anterior facial

height. Thus, when considering the vertical dimension,
NA

NA

a major issue during treatment planning that has also

(mean 6 SD), sex

gained much interest in the literature is the decision

age, 12.4 6 1.1 y

age, 18.5 6 3.6 y

Sample size, age
Intervention

Nonextraction

Nonextraction

Nonextraction

whether to extract teeth. This is a highly controversial

n 5 23 (12 F,

n 5 29 (18 F,
treatment

treatment

treatment

Age, 12.8 y

issue, since certain studies support the claim that pos-

11 M)

11 M)

n 5 45

terior tooth extractions reduce or, for hyperdivergent

patients, better control the vertical dimension
compared with nonextraction strategies.2,6,7 However,
for anchorage in
age, 12.3 6 1.0 y high-pull headgear

NA, Not available; M, male; F, females; Class II, 1, Class II Subdivision 1.

intermaxillary

this systematic review clearly demonstrates no

hyperdivergent n 5 23 (14 F, 9 M) elastics, and
Treatment

all patients
details
Changes in vertical Four first premolar Lip bumper,

available evidence to support this clinical belief.

Extraction group (Ex)

Although various studies in this review reported in

their conclusions sections that the decision for
Changes in vertical Four first premolar NA

Changes in vertical Four first premolar NA

extractions affected the vertical dimension, this was

Class I patients n 5 31 (26 F, 5 M)
Main objective (mean 6 SD), sex
Sample size, age

age, 17.2 6 3.9 y

not supported by their data.17,18 This fact indicates

Intervention

extractions

extractions

extractions

how deeply rooted is this perception in clinicians’

age, 14.4 y

beliefs.
Class I patients n 5 38

We tested the effect of 4 premolar extractions, which

is the most commonly used approach, for the skeletal
vertical dimension. The 14 included studies contained
dimension of

dimension of

dimension of

samples of patients who had different malocclusions,

patients

vertical growth patterns, and developmental statuses.

Table I. Continued

No study clearly confirmed the expected effect: ie, a

reduction or at least a smaller increase in the vertical
dimension of patients who had 4 premolar extraction
Sivakumar and
Valiathan18
Retrospective

Retrospective

Retrospective
Study design
Porto et al17

treatment, compared with nonextraction treatment.

Staggers19

Moreover, apart from 1 study20 that showed a 2

Author

(2012)

(2008)

(1994)
Year

decrease in FMA in the nonextraction group, no study

showed a mean change in the angular measurements

American Journal of Orthodontics and Dentofacial Orthopedics August 2018  Vol 154  Issue 2
August 2018  Vol 154  Issue 2

182
Table II. Risk of bias of the included studies
Bias in measurement Bias due to deviations
Selection classification of from intended Bias due to Bias in measurement Bias in selection of
Study Confounding bias interventions interventions missing data of outcomes reported result Overall
Basciftci and Class I: Moderate Moderate No information Low Moderate Moderate Serious
Usumez23 moderate (extracted premolars (NA treatment details) (not blinded assessor)
(2003) not defined)
Class II: critical Moderate Moderate No information (NA Low Moderate Moderate Critical
(difference in age and (extracted premolars treatment details) (not blinded assessor)
SN-GoGn pretreatment) not defined)
Bravo et al22 Moderate Moderate Moderate No information (NA Low Moderate Moderate Serious
(1997) (2 extraction patterns treatment details) (not blinded assessor)
were used)
Gkantidis et al8 Moderate Low Low Low Low Moderate Low Moderate
(2011) (not blinded assessor)
Hayasaki et al9 Moderate Moderate Low Low Low Moderate Low Moderate
(2005) (not blinded assessor)
Hosseinzadeh-Nik Moderate Moderate Low No information Low Moderate Moderate Serious
et al21 (NA treatment details) (not blinded assessor)
(2016)
American Journal of Orthodontics and Dentofacial Orthopedics

Katsaros et al24 Moderate Moderate Moderate No information (NA Low Moderate Low Serious
(1996) (extracted premolars treatment details) (not blinded assessor)
not defined)
Kirschneck et al13 Moderate Low Low Moderate (treatment details Low Low Low Moderate
(2015) partially provided)
Kocadereli14 Moderate Moderate Low Moderate Low Moderate Low Moderate
(1999) (treatment details partially (not blinded assessor)
provided)
Kumari and Fida15 Serious Moderate Low Moderate Low Serious Serious Serious
(2010) (unknown treatment time (treatment details partially (no method error, not (only posttreatment
and sex distribution) provided) blinded assessor) values compared)
Luppanapornlap and Critical Low Low No information (NA Low Moderate Low Critical
Johnston16 (different group baseline treatment details) (not blinded assessor)
(1993) characteristics)
Paquette et al20 Moderate Low Low No information (NA Low Moderate Low Serious
(1992) treatment details) (not blinded assessor)
Porto et al17 Moderate Moderate Low Low Low Moderate Low Moderate
(2012) (not blinded assessor)
Sivakumar and Critical Moderate Low No information (NA Low Moderate Low Critical
Valiathan18 (differences in treatment treatment details) (not blinded assessor)

Kouvelis et al
(2008) time and sex distribution)
Staggers19 No Information Moderate Low No information (NA Low Serious Low Critical
(1994) (baseline table missing, treatment details) (NA method error, not
unknown sex distribution) blinded assessor)

NA, Not available.

Kouvelis et al 183

Table III. Interventions and outcomes of the included studies

Author
Year Results
Study design Intervention Mean 6 SD of T1-T0 difference Effect on vertical dimension
Basciftci and Class I Class I There is no difference between extraction and
Usumez23 4 premolar Ex (NA) vs Nonex SN-GoGn ( ), Ex:
1.0 6 3.2; nonextraction treatment, in both dental Class I
(2003) Class II, 1 Nonex:
0.4 6 2.3; P 5 0.460 and Class II, 1 hyperdivergent patients.
Retrospective 4 premolar Ex (NA) vs Nonex Class II, 1
SN-GoGn ( ), Ex: 10.2 6 2.5;
Nonex: 10.8 6 2.0; P 5 0.425
Bravo et al22 4 first premolar or 2 maxillary FMA ( ), Ex:
0.6 6 1.7; Nonex: There is no difference between extraction and
(1997) first and 2 mandibular
0.4 6 1.4; P 5 0.724 nonextraction treatment in dental Class II
Retrospective second premolar Ex vs normodivergent patients.
Nonex
Gkantidis et al8 4 first premolar Ex vs Nonex SN-GoGn ( ), Ex:
0.1 6 2.5; There is no difference between extraction and
(2011) Nonex: 10.1 6 1.7; P 5 0.616 nonextraction treatment in dental Class II, 1
Retrospective FH-MP ( ), Ex:
0.2 6 2.5; hyperdivergent patients.
Nonex: 10.1 6 2.4; P 5 0.709
ANS-Me (mm), Ex: 14.0 6 4.50
Nonex: 13.3 6 3.1; P 5 0.643
LFH/TFH, Ex: 10.1 6 0.9;
Nonex: 10.2 6 1.5; P 5 0.822
Hayasaki et al9 Class I Class I There is no difference between extraction and
(2005) 4 first premolar Ex vs Nonex SN-GoGn ( ), Ex:
0.8 6 2.0; nonextraction treatment in dental Class I and
Retrospective Class II, 1 Nonex: 10.8 6 2.6; P 5 0.071 Class II, 1 normodivergent patients.
4 first premolar Ex vs Nonex Me-PP (mm), Ex: 12.8 6 2.6;
Nonex: 13.5 6 2.7; P 5 0.456
N-Me (mm), Ex: 14.5 6 3.8,
Nonex: 16.0 6 4.4; P 5 0.341
ANS-Me (mm), Ex: 12.4 6 2.4;
Nonex: 13.3 6 2.8; P 5 0.385
Class II
SN-GoGn ( ), Ex:
1.1 6 2.5;
Nonex:
0.4 6 2.1; P 5 0.446
Me-PP (mm),Ex: 12.6 6 2.3;
Nonex: 13.5 6 2.1; P 5 0.263
N-Me (mm), Ex: 14.6 6 2.9;
Nonex: 15.7 6 4.3; P 5 0.445
ANS-Me (mm), Ex: 12.3 6 2.3;
Nonex: 13.3 6 2.0; P 5 0.217
Hosseinzadeh-Nik 4 first premolar Ex vs Nonex Pog-SN7 (mm), Ex:
0.4 6 9.8; There is no difference between extraction and
et al21 Nonex:
1.0 6 6.9; P 5 0.833 nonextraction treatment in dental Class I
(2016) B-SN7 (mm), Ex: 12.0 6 8.3; bimaxillary protrusion patients.
Retrospective Nonex:
0.4 6 6.5; P 5 0.341
Katsaros et al24 4 premolar Ex (NA) vs Nonex SN-GoMe ( ), Ex:
0.8 6 2.7; There is no difference between extraction and
(1996) Nonex:
0.3 6 2.2; P 5 0.431 nonextraction treatment in dental Class II, 1
Retrospective PP-GoME ( ), Ex:
1.2 6 2.9; hyperdivergent patients.
Nonex:
1.1 6 3.0; P 5 0.894
Kirschneck et al13 4 first premolar Ex vs Nonex SN-GoGn ( ), Ex:
0.9 6 2.3; There is no difference between extraction and
(2015) Nonex:
1.1 6 2.5; P 5 0.770 nonextraction treatment in hyperdivergent
Retrospective PP-GoGn ( ), Ex:
1.2 6 2.7; patients of varying malocclusion.
Nonex:
0.9 6 2.4; P 5 0.680
Kocadereli14 4 first premolar Ex vs Nonex SN-GoGn ( ), Ex: 10.1 6 3.5; There is no difference between extraction and
(1999) Nonex: 10.3 6 2.8; P 5 0.790 nonextraction treatment in dental Class I
Retrospective FMA ( ), Ex:
0.4 6 3.2; hyperdivergent patients.
Nonex: 10.1 6 3.0; P 5 0.412
N-Me (mm), Ex: 14.7 6 5.0;
Nonex: 16.0 6 5.1; P 5 0.234
ANS-Me (mm), Ex: 12.9 6 3.2;
Nonex: 13.8 6 3.7; P 5 0.292

American Journal of Orthodontics and Dentofacial Orthopedics August 2018  Vol 154  Issue 2
184 Kouvelis et al

Table III. Continued

Author
Year Results
Study design Intervention Mean 6 SD of T1-T0 difference Effect on vertical dimension

Kumari and Fida 15
4 first premolar Ex FMA ( ), Ex:
0.2; Nonex: 10.6; P .0.050 There is a statistically significant increase in
(2010) vs Nonex N-Me (mm), Ex: 11.5; Nonex: 15.5; N-Me distance (4 mm) in the nonextraction
Retrospective P \0.05* compared with extraction treatment in skeletal
N-ANS/ANS-Me, Ex: 0.0, Nonex: 11.1; Class I hyperdivergent patients; no difference
P .0.050 was detected in FMA angle and N-ANS/ANS-
Me ratio.
Luppanapornlap and 4 first premolar Ex vs Nonex FMA ( ), Ex:
0.2, Nonex: 10.7; P .0.050 There is no difference between extraction and
Johnston16 N-Me (mm), Ex: 15.4; Nonex: 15.0; nonextraction treatment in dental Class II, 1
(1993) P .0.050 patients.
Retrospective ANS-Me (mm), Ex: 12.7, Nonex: 13.4;
P .0.050
Paquette et al20 4 first premolar Ex vs Nonex FMA ( ), Ex: 10.3; Nonex:
2.0; There is a statistically significant, but clinically
(1992) P \0.010* questionable decrease in FMA angle (2.3 ) in
Retrospective N-Me (mm), Ex: 15.4; Nonex: 17.0; the nonextraction compared to extraction
P .0.050 treatment in dental Class II, 1 normodivergent
ANS-Me (mm), Ex: 13.0, Nonex: 13.9; patients; no difference was detected in
P .0.050 N-Me or ANS-Me distance.
Porto et al17 4 first premolar Ex vs Nonex SN-GoGn ( ), Ex:
0.9 6 2.4; There is a statistically significant, but clinically
(2012) Nonex: 10.8 6 2.3; P 5 0.020* questionable increase in SN-GoGn angle (1.7 )
Retrospective FMA ( ), Ex:
0.3 6 3.1; in nonextraction compared with extraction
Nonex: 10.7 6 2.5; P 5 0.240 treatment in dental Class II, 1 hyperdivergent
patients; no difference was detected in FMA
angle.
Sivakumar and 4 first premolar Ex vs Nonex FMA ( ), Ex: 10.5 6 1.7; There is a statistically significant, but clinically
Valiathan18 Nonex: 10.1 6 2.0; P 5 0.985 questionable increase in N-Me distance
(2008) N-Me (mm), Ex: 12.3 6 3.0; (1.4 mm) in the extraction compared with
Retrospective Nonex: 10.9 6 2.1; P 5 0.033* nonextraction treatment in dental Class I
ANS-Me (mm), Ex: 11.3 6 1.3; normodivergent patients; no difference was
Nonex: 10.6 6 1.8; P 5 0.080 detected in FMA angle and ANS-Me distance.
Staggers19 4 first premolar Ex vs Nonex MP-HP ( ), Ex: 10.1 6 2.5; There is no difference between extraction and
(1994) Nonex: 10.1 6 2.0; P 5 0.950 nonextraction groups in dental Class I
Retrospective N-Me (mm), Ex: 15.4 6 4.7; patients.
Nonex: 15.1 6 3.6; P 5 0.740
N-ANS/ANS-Me, Ex:
1.4 6 3.8; Nonex:

2.7 6 3.2; P 5 0.090
NA, Not available; Ex, extraction; Nonex, nonextraction; Class II, 1, Class II Subdivision 1.
*P \0.05.

greater than 1.2 in either extraction or nonextraction
group. Thus, these changes cannot be considered to
have clinical significance.
The main rationale behind the supposition that tooth
extractions reduce the vertical dimension is based on the
wedge-effect concept, according to which the mesial
movement of the molars during closure of the extraction
spaces reduces the vertical dimension by anterior rota-
tion of the mandible.25 However, this concept has not
been proven. A possible explanation for this, as reported
by Gkantidis et al,8 is that the direction of the mesializa-
tion of the posterior teeth during space closure is parallel
to the occlusal plane, and not to the maxillary and
mandibular planes for the maxillary and mandibular
teeth, respectively, as the wedge-effect theory suggests.
The craniofacial complex, including the masticatory sys-
tem, is complicated and should not be perceived as a
simple articulator. There are probably more important
factors than tooth numbers responsible for the establish-
ment of the vertical positions of the teeth and the asso-
ciated skeletal patterns, such as neuromuscular balance
and function.8
For the patterns of extractions, various combinations
have been reported in the literature and are used in clinical
practice; the most popular is 4 first or second premolar
extraction treatment. For this reason, our review focused
only on these patterns. A preliminary literature search iden-
tified studies that tested the extraction of 4 first or
second molars7,19,26 or 2 mandibular premolars,27 but these
did not have a treated nonextraction control group.

August 2018  Vol 154  Issue 2 American Journal of Orthodontics and Dentofacial Orthopedics
Kouvelis et al
Most studies in this review assessed hyperdivergent
rather than normodivergent patients, and no study had
an extraction group consisting of only hypodivergent
patients, supporting the widely accepted clinical belief
that a hypodivergent profile is a contraindication for ex-
tractions, whereas hyperdivergency is an indication for
extractions. In contrast to this, we found no difference
in normodivergent and hyperdivergent patients. Only
Kumari and Fida15 found a clinically significant increase
in N-Me distance (4 mm) in nonextraction compared
with extraction treatment, but the study had a serious
risk of bias and inadequate reporting of the results.
Furthermore, it did not detect a difference in FMA.
Considering the growth status, most studies included
both growing and nongrowing patients. Some studies
attempted to control the confounding due to growth
status, age, and sex by including only female pa-
tients13,22,24 or having a similar sex distribution
between the compared groups.8,9,17,23 Three studies
included only nongrowing patients.15,18,21
Regarding malocclusions, most studies included pa-
tients with Class I or Class II (primarily Division 1) maloc-
clusion. Class III patients were included only in the study
of Kirschneck et al,13 but they investigated a balanced
sample of patients with various malocclusion types
and did not report separate results for each malocclusion
class.
Thus, based on the available evidence, which is a
considerable amount but has low to moderate quality,
4 premolar extraction treatment does not seem to have
any influence on the patient's skeletal vertical dimen-
sion, regardless of age, sex, malocclusion type, or vertical
growth pattern.
Limitations
A major limitation of this review proved to be the
problematic reporting. For example, factors and vari-
ables important for the assessment of comparability
among studies, such as malocclusion pattern, type of ex-
tracted teeth, growth status, or exact treatment details,
were often not adequately reported. Furthermore,
certain studies reported no correction of the magnifica-
tion factor of the lateral cephalograms when assessing
linear measurements, thus adding inaccuracies to these
variables. As a result, poor reporting combined with
the considerable methodological, statistical, and clinical
heterogeneity among studies precluded any quantitative
synthesis.
This review was limited to 4 premolar extraction
treatments, because the few controlled studies in the
literature that investigated different extraction patterns
did not include a treated control group. Furthermore,
American Journal of Orthodontics and Dentofacial Orthopedics
185
all included studies were retrospective, and this is an
inherent limitation of this review, since retrospective
studies are prone to various types of biases and especially
to selection bias. The lack of prospective or randomized
clinical trials might be attributed to the difficulty in col-
lecting these samples prospectively, for both practical
and ethical concerns.
CONCLUSIONS
In this systematic review, it was reasonable to
conclude that there is no specific effect of 4 premolar
extraction treatment compared with nonextraction
treatment protocols on the vertical dimension of the
face. Despite the low to moderate quality of evidence,
there is considerable agreement among the available
studies that an extraction protocol to reduce or control
vertical dimension is not justified. Although randomized
controlled trials may not be feasible for such treatment
options, well-conducted prospective cohort studies
would be desirable to confirm this statement.
REFERENCES
1. Schudy FF. Vertical growth versus anteroposterior growth as
related to human function and treatment. Angle Orthod 1964;
34:75-93.
2. Schudy FF. The control of vertical overbite in clinical orthodontics.
Angle Orthod 1968;38:19-39.
3. Isaacson JR, Isaacson RJ, Speidel TM, Worms FW. Extreme varia-
tion in vertical facial growth and associated variation in skeletal
and dental relations. Angle Orthod 1971;41:219-29.
4. Fr€ankel R, Fr€ankel C. A functional approach to treatment of skel-
etal open bite. Am J Orthod 1983;84:54-68.
5. Pearson LE. Vertical control in treatment of patients having
backward-rotational growth tendencies. Angle Orthod 1978;48:
132-40.
6. Garlington M, Logan LR. Vertical changes in high mandibular
plane cases following enucleation of second premolars. Angle Or-
thod 1990;60:263-8.
7. Aras A. Vertical changes following orthodontic extraction treat-
ment in skeletal open bite subjects. Eur J Orthod 2002;24:407-16.
8. Gkantidis N, Halazonetis DJ, Alexandropoulos E, Haralabakis NB.
Treatment strategies for patients with hyperdivergent Class II Divi-
sion 1 malocclusion: is vertical dimension affected? Am J Orthod
Dentofacial Orthop 2011;140:346-55.
9. Hayasaki SM, Castanha Henriques JF, Janson G, de Freitas MR. In-
fluence of extraction and nonextraction orthodontic treatment in
Japanese-Brazilians with Class I and Class II Division 1 malocclu-
sions. Am J Orthod Dentofacial Orthop 2005;127:30-6.
10. Bansal A. Long face syndrome: a literature review. J Dent Health
Oral Disord Ther 2015;2:1-6.
11. Deeks J, Higgins J, Altman D. Analysing data and undertaking
meta-analyses. In: Higgins J, Sally G, editors. Cochrane handbook
for systematic reviews of interventions. 5th ed. London, United
Kingdom: Cochrane Collaboration; 2011.
12. Sterne JA, Hernan MA, Reeves BC, Savovic J, Berkman ND,
Viswanathan M, et al. ROBINS-I: a tool for assessing risk of
bias in non-randomised studies of interventions. BMJ 2016;355:
i4919.
August 2018  Vol 154  Issue 2
186
13. Kirschneck C, Proff P, Reicheneder C, Lippold C. Short-term effects
of systematic premolar extraction on lip profile, vertical dimension
and cephalometric parameters in borderline patients for extraction
therapy—a retrospective cohort study. Clin Oral Investig 2015;20:
865-74.
14. Kocadereli I. The effect of first premolar extraction on vertical
dimension. Am J Orthod Dentofacial Orthop 1999;116:41-5.
15. Kumari M, Fida M. Vertical facial and dental arch dimensional
changes in extraction vs. non-extraction orthodontic treatment.
J Coll Physicians Surg Pak 2010;20:17-21.
16. Luppanapornlarp S, Johnston LE. The effects of premolar extrac-
tion: a long-term comparison of outcomes in “clear-cut” extrac-
tion and nonextraction Class II patients. Angle Orthod 1993;63:
257-72.
17. Porto VS, Henriques JF, Janson G, Freitas MR, Pinzan A. Influence
of treatment with and without extractions on the growth pattern
of dolichofacial patients. Dental Press J Orthod 2012;17:69-75.
18. Sivakumar A, Valiathan A. Cephalometric assessment of dentofa-
cial vertical changes in Class I subjects treated with and without
extraction. Am J Orthod Dentofacial Orthop 2008;133:869-75.
19. Staggers JA. Vertical changes following first premolar extractions.
Am J Orthod Dentofacial Orthop 1994;105:19-24.
20. Paquette DE, Beattie JR, Johnston LE. A long-term comparison of
nonextraction and premolar extraction edgewise therapy in
August 2018  Vol 154  Issue 2 American Journal of Orthodontics and Dentofacial Orthopedics
Kouvelis et al
“borderline” Class II patients. Am J Orthod Dentofacial Orthop
1992;102:1-14.
21. Hosseinzadeh-Nik T, Eftekhari A, Shahroudi AS, Kharrazifard MJ.
Changes of the mandible after orthodontic treatment with and
without extraction of four premolars. J Dent (Tehran) 2016;13:
199-206.
22. Bravo LA, Canut JA, Pascual A, Bravo B. Comparison of the
changes in facial profile after orthodontic treatment, with and
without extractions. Br J Orthod 1997;24:25-34.
23. Basciftci FA, Usumez S. Effects of extraction and nonextraction
treatment on class I and class II subjects. Angle Orthod 2003;73:
36-42.
24. Katsaros C, Ripplinger B, H€ogel A, Berg R. Der Einfluß der kiefer-
orthop€adischen Extraktions- bzw. Nichtextraktionsbehandlung
auf das Weichteilprofil. J Orofac Orthop 1996;57:354-65.
25. Fields HW, Proffit WR, Nixon WL, Phillips C, Stanek E. Facial
pattern differences in long-faced children and adults. Am J Orthod
1984;85:217-23.
26. Hans MG, Groisser G, Damon C, Amberman D, Nelson S,
Palomo JM. Cephalometric changes in overbite and vertical facial
height after removal of 4 first molars or first premolars. Am J Or-
thod Dentofacial Orthop 2006;130:183-8.
27. Al-Nimri KS. Vertical changes in class II division 1 malocclusion af-
ter premolar extractions. Angle Orthod 2006;76:52-8.
Kouvelis et al 187

APPENDIX

MEDLINE SEARCH STRATEGY

1 Tooth extraction 22826
2 extraction AND orthodont* 5002
3 tooth extraction AND orthodont* AND vertical 331
4 extract*[tiab] AND orthodont*[tiab] AND vertical[tiab] 266
5 (vertical[Title/Abstract] OR caudal[Title/Abstract] OR 287
cranial[Title/Abstract]) AND (tooth extraction AND
orthodont*)
6 (tooth extraction[Title/Abstract] AND orthodont*[Title/ 19
Abstract]) AND (vertical[Title/Abstract] OR caudal
[Title/Abstract] OR cranial[Title/Abstract])
7 (extract*[Title/Abstract] AND orthodont*[Title/Abstract]) 17
AND perpendicular[Title/Abstract]
8 ((premolar*[Title/Abstract] AND extract*[Title/Abstract]) 137
AND (vertical[Title/Abstract] OR caudal[Title/Abstract]
OR cranial[Title/Abstract])) AND orthodont*[Title/
Abstract]
9 ((premolar* OR molar*[Title/Abstract]) AND extract* 193
[Title/Abstract]) AND (vertical[Title/Abstract] OR
caudal[Title/Abstract] OR cranial[Title/Abstract]) AND
orthodont*[Title/Abstract]
10 (tooth extraction[Title/Abstract] AND orthodont*[Title/ 13
Abstract]) AND cephalometr*[Title/Abstract]
11 tooth extraction[MeSH Major Topic] AND orthodont* 47
[Title/Abstract] AND vertical[Title/Abstract]

American Journal of Orthodontics and Dentofacial Orthopedics August 2018  Vol 154  Issue 2