A Systematic Review and Meta-Analysis

886322 JOO Journal of OrthodonticsMachado et al.
Scientific Section
Journal of Orthodontics
A systematic review and meta-analysis 2020, Vol. 47(1) 7–29

https://doi.org/10.1177/1465312519886322
DOI: 10.1177/1465312519886322
© The Author(s) 2019
on Bolton’s ratios: Normal occlusion Article reuse guidelines:
sagepub.com/journals-permissions
and malocclusion journals.sagepub.com/home/joo
Vanessa Machado1,2 , João Botelho1, Paulo Mascarenhas1,

José João Mendes1 and Ana Delgado1,2
Abstract
Introduction: The purpose of this study was to seek and summarise the Bolton overall index (OI) and anterior index
(AI) regarding normal occlusion and Angle’s malocclusion according to gender, and to assess if these indices support
Bolton’s standards as general references.
Methods: PubMed, LILACS, Embase, CENTRAL and Google Scholar databases were searched up to June 2019
(CRD42018088438). Non-randomised clinical studies, published in English and assessing Bolton’s OI and/or AI in normal
occlusion and Angle’s malocclusion groups, were included. OI and AI means, sample size and SDs were collected. The
National Heart, Lung, and Blood Institute’s Quality Assessment Tool for Observational Cohort and Cross-Sectional
Studies was used to assess the risk of bias. Pairwise random-effects and multilevel Bayesian network meta-analyses were
used to synthesise available data.
Results: Fifty-three observational studies were included (11,411 participants; 3746 men, 4430 women; 15 studies lacked
gender information). For normal occlusion, pooled estimates for OI and AI means were 91.78% (95% confidence interval
[CI] = 91.42–92.14; I2 = 92.87%) and 78.25% (95% CI = 77.87–78.62; I2 = 90.67%), respectively. We could identify in
Angle’s Class III patients meaningful OI and AI mean deviations from normal occlusion (0.76, 95% credible interval [CrI]
= 0.55–0.98 and 0.61, 95% CrI = 0.35–0.87, respectively), while in Class II patients we found a meaningful mean deviation
from normal occlusion only for OI (−0.28, 95% CrI = −0.52–−0.05). Concerning gender impact, male patients presented
higher OI (0.30, 95% CI = 0.00–0.59) and AI (0.41, 95% CI = 0.00–0.83) mean values than female patients in Class I.
Conclusion: Normal occlusion OI and AI mean values differ from Bolton’s original values. Class II division 2, for OI
mean values, and Class III, for both OI and AI, are proportionally larger than normal occlusion patients. Gender had
almost no impact on teeth mesiodistal proportion.
Keywords
tooth size, tooth size discrepancy, Bolton ratios, meta-analysis, systematic review
Date received: 16 February 2019; revised: 14 September 2019; accepted: 13 October 2019
Introduction 1Clinical Research Unit (CRU), Centro de Investigação Interdisciplinar

Egas Moniz (CiiEM), Egas Moniz - Cooperativa de Ensino Superior,
An appropriate balance of mesiodistal tooth widths between C.R.L., Monte de Caparica, Almada, Portugal
2Orthodontics Department, Clinical Research Unit (CRU), Centro de
the maxillary and mandibular arches allows correct inter-
Investigação Interdisciplinar Egas Moniz (CiiEM), Egas Moniz - Cooperativa
digitation, overbite and overjet in normal occlusion, with de Ensino Superior, C.R.L., Monte de Caparica, Almada, Portugal
proper aesthetic and function (Bolton, 1958). Currently, the
extent of mesiodistal movement grounds clinical practice Corresponding author:
from conventional through to orthodontic aligner treat- Vanessa Machado, Clinical Research Unit (CRU), Centro de Investigação
Interdisciplinar Egas Moniz (CiiEM), Egas Moniz - Cooperativa de Ensino
ments. It has also allured clinical interest, particularly in Superior, C.R.L., Campus Universitário, Quinta da Granja, Monte de
anteroposterior malocclusions correction (Kravitz et al., Caparica, Caparica, 2829-511, Portugal.
2009; Lombardo et al., 2017). Email: vmachado@egasmoniz.edu.pt
8 Journal of Orthodontics 47(1)
The concept of a proportional balance between the mesi- and malocclusions in multiple populations (Endo et al., 2008;
odistal sums of maxillary and mandibular teeth may have Kumar et al., 2013; Manopatanakul and Watanawirun, 2011;
arisen from the beginnings of dental articulation theories. O’Mahony et al., 2011; Ricci et al., 2013; Santoro et al., 2000),
Bonwill (1899) stated that ‘Nature left to herself, always there is no consensus about its correlation with the different
brings proposition. . . the proportions of upper teeth to the types of malocclusions classified by Angle.
lower teeth are as exact as any’. This nature theory was per-
vasive in early orthodontics and was seen in the strict non- Objectives
extraction period started by Edward Angle. The mesiodistal
widths of teeth were initially investigated by Black (1902). No study has investigated, in an evidence-based manner,
Historically, Young (1923) was the first to study the inter- normative values for mesiodistal proportions of normal
maxillary tooth width ratio in occlusion and, thereafter, occlusion and Angle’s malocclusion, from worldwide
Gilpatric (1923) found that the upper arch was 8–12 mm researched data. For that reason, the primary aim of this sys-
wider than the lower arch. Over the years, to account and be tematic review was to synthesise worldwide estimates for
aware of this proportion, several methods have been sug- normal occlusion OI and AI mean values, and to compare
gested to assess the interarch tooth size relationship (Bolton, such values with those proposed by Bolton, to address the
1968, 1962; Kesling, 1945; Neff, 1949, 1957), but Bolton’s following focused question: Are current standards globally
ratios have become widely applied in orthodontics research. appropriate? Second, we aimed to compare the obtained val-
In this regard, the overall index (OI) is the percentage ues between gender for normal occlusion and each malocclu-
obtained by summing the widths of the 12 mandibular teeth sion type. Finally, we intended to obtain OI and AI pooled
divided by the sum of the widths of the 12 maxillary teeth. estimates for each type of Angle’s malocclusion and com-
The anterior index (AI) is the percentage obtained by sum- pare them against the obtained values for normal occlusion
ming the widths of the six mandibular anterior teeth divided under a multilevel Bayesian network meta-analysis model.
by the sum of the widths of the six maxillary anterior teeth
(Bolton, 1958, 1962). On average, the OI was 91.3% (± Materials and methods
1.91) and AI was 77.2% (± 1.65); these promptly became
standard values in the diagnosis and guidance of orthodon- Protocol and registration
tic treatments. The protocol for this systematic review was made a priori,
Over time, Bolton’s analyses have proved to be clini- agreed upon by all authors and registered in PROSPERO
cally useful in extreme teeth size discrepancies. However, (ID number: CRD42018088438). This systematic review
without neglecting its value, its methodology and conclu- was conducted according to the Cochrane Handbook
sions should be carefully evaluated. First, these studies had (Higgins et al., 2003) and reported according to the
a potential selection bias since the population was not spec- PRISMA statement (Preferred Reporting Items for
ified, particularly concerning race, ethnicity and gender. Systematic Reviews and Meta-Analyses) (Liberati et al.,
Second, although the author has stated that his ratios were 2009) (Supplement S1) and its extension for abstracts
based on 55 cases ‘where excellent occlusions existed’, 44 (Beller et al., 2013).
models were from patients who underwent orthodontic
treatment and only 11 were untreated (Bolton, 1958).
Eligibility criteria
According to the literature, teeth size variation is ethnic-
and gender-related (Bishara et al., 1989; Black, 1902; Studies were eligible for inclusion based on the following
Hattab et al., 1996; Lavelle, 1972; Santoro et al., 2000; criteria:
Smith et al., 2000), pointing out an anthropological signifi-
cance with genetic underpinnings (Dempsey and Townsend, 1. Randomised and non-randomised (cohort/longitudi-
2001; Hughes et al., 2000). For this reason, the application nal or cross-sectional studies);
of Bolton analyses and the proposed standard values for a 2. English language studies;
harmonious dentition might not be valid for other popula- 3. Human study population;
tions. Therefore, this population-based variation has 4. Determined Bolton’s analysis with normal occlusion
become a subject of interest for many researchers, resulting and/or Angle’s Class I, Class II, Class II division 1,
in several attempts to establish normative standards for dif- Class II division 2 and/or Class III, in patients with-
ferent racial groups (Al-Khateeb and Abu Alhaija, 2006; out previous orthodontic treatment;
Al-Omari et al., 2008; Bernabé et al., 2004; Lavelle, 1972; 5. Dental casts or digital models with all permanent
Santoro et al., 2000). teeth from the maxillary and mandibular right first
Another relevant question is the relationship between the molar to the left first molar completely erupted,
tooth size discrepancy for both OI and AI and the various types without tooth deformities (Scheid and Weiss, 2012),
of Angle’s malocclusion. Although several investigators have mesiodistal restorations, caries or abrasion that
emphasised the relationship between Bolton ratio discrepancies could affect the teeth’s mesiodistal diameter;
Machado et al. 9
6. The study measured the largest mesiodistal teeth Risk of bias

dimension to the nearest 0.01 mm, through digital
calliper or software. The Quality Assessment Tool for Observational Cohort and
Cross-Sectional Studies statement proposed by National
Narrative reviews, case reports and case series studies were Heart, Lung, and Blood Institute (NIHLBI) was used to
excluded from review. appraise study quality (https://www.nhlbi.nih.gov/health-
topics/study-quality-assessment-tools). The checklist was
adapted since criteria 7, 8, 10 and 13 did not apply. The
Search strategy reviewers (VM, JB) determined a total quality score for
A systematic search was conducted and updated in June each article. Each methodologic quality criterion was
2019, covering the following electronic databases: PubMed, assigned 1 point, to a total maximum of 10 achievable
LILACS, Embase, CENTRAL (The Cochrane Central points. Studies reaching 9 or 10 points were arbitrarily con-
Register of Controlled Trials) and Google Scholar. The sidered of high quality, studies with 7 or 8 points were clas-
strategy used for the electronic search was the following: sified as medium quality, and studies with 6 points or less
[‘Bolton ratio’ OR ‘tooth size discrepancy’ OR ‘Bolton dis- considered of low methodologic quality. To be included,
crepancy’ OR ‘tooth-size ratios’ OR ‘tooth-size measure- articles could not be of low quality, as recommended by the
ment’ OR ‘Bolton analysis’]. Cochrane Handbook (Higgins et al., 2003).
No limitations were applied regarding publication
year. The reference lists of included articles and relevant
reviews were manually searched. Grey literature was
Summary measures and synthesis of results
searched using the latter strategy in OpenGrey. Authors The objective of synthesis of the initial (priors) normal
were contacted when necessary for additional data or occlusion OI and AI mean values was accomplished by pair-
clarifications. wise random effects meta-analysis using OpenMetaAnalyst
(Wallace et al., 2017) software. Quantities I2 and Tau2 were
measured to account for the heterogeneity associated with
Assessment of eligibility
the Bolton ratios mean estimates. Funnel plots were used to
The eligibility of each study was assessed independently visualise and quantify meta-analysis publication bias, respec-
by two investigators (VM and JB) who screened the titles tively, if appropriate (Doi et al., 2015a, 2015b; Egger and
and/or abstracts of retrieved studies. Inclusion was depend- Smith, 1998; Furuya-Kanamori et al., 2018; Higgins and
ent on the following eligibility criteria: randomised or non- Green, 2011; Sterne et al., 2011). All tests were two-tailed
randomised studies with OI and/or AI data. Final selection with alpha set at 0.05 except for the z-test whose significance
of studies was performed by three authors independently level cut-off was adjusted to 0.10 when comparing meta-
(JB, VM, PM) and verified by a fourth and fifth author analysis outcomes. Unpaired z-test was used to compare our
(JJM, AD), by reviewing the full text based on the inclu- normal occlusion mean results with Bolton original values.
sion criteria above. Discussion resolved any disagree- In a number of articles (Asma, 2013; Bugaighis et al.,
ments. Non-full papers, such as conference abstracts and 2015; Crosby and Alexander, 1989; Kansal et al., 2012;
letters to editors, were excluded. Machado et al., 2018; Mahmoud et al., 2017; McSwiney
et al., 2014; O’Mahony et al., 2011; Oktay and Ulukaya,
2010; Uysal et al., 2005), Class II division 1 and division 2
Data extraction
summary statistics were published separately, and it was
Data were extracted to a predefined table. We used the necessary to calculate the combined mean and SD for the
following information: the first author’s name; study overall Class II following the algorithms in Altman et al.
design; publication year; country and continent where the (2000). Similarly, in the studies by McSwiney et al. (2014)
study was conducted; number of cases and participants; and Nie and Lin (1999), there were published data for surgi-
gender; tooth width measurement method; and OI and AI cal and non-surgical in Class III, and we used the aforemen-
(mean and SD). Type of occlusion was classified into tioned procedure to combine the mean and SD.
normal occlusion, Angle’s Class I, Class II (division 1 To fulfil the second objective, we were required to esti-
and division 2) or Class III. Populations were categorised mate the OI and AI overall mean of Angle’s malocclusion
into continental groups: African; American; Asia (includ- types and compare each of them against the normal occlu-
ing Japanese populations based in Hawaii); European; sion overall mean value. Due to the complexity of the distri-
and Oceania. We extracted Bolton OI, AI means and SDs, bution of the extracted data across the selected reports, we
for both genders, in all selected studies population aimed to address such an issue under a multilevel Bayesian
samples. Concerning additional data/clarifications, we
network meta-analysis (Bayesian NMA) model, as previ-
tried to contact corresponding authors (until June 2019) ously used (Barbato et al., 2015; Kotsakis et al., 2018;
(Supplement S2). Tedesco et al., 2018). Bayesian NMA has been developed to
cope with limitations in traditional pairwise meta-analysis. Study characteristics

NMA incorporates all available evidence into a general
framework model for comparisons of all available factors. A Setting. Table 1 summarises the characteristics of the
further development in the NMA is to use a multilevel included studies. In total, the analysis included 11,411
Bayesian statistical approach, which provides a more flexi- participants (3746 men, 4430 women). However, 16 stud-
ble modelling framework to take into account heterogeneity ies (Al-Duliamy et al., 2016; Alkofide and Hashim, 2002;
in the evidence and covariance between means due to the Asma, 2013; Bugaighis et al., 2015; Cançado et al., 2015;
complexity in the data structure. Therefore, we went through Chugh et al., 2015; Crosby and Alexander, 1989; Kansal
Rstan, a R package that inter-connects R and STAN lan- et al., 2012; Mahmoud et al., 2017; Manopatanakul and
guages making the design and handling of such a complex Watanawirun, 2011; McSwiney et al., 2014; Mulimani
model easy. Priors were sourced as the initial mean and vari- et al., 2018; Ricci et al., 2013; Shastri et al., 2015; Zer-
ance for AI and OI normal occlusion values from the pair- ouaoui et al., 2014) lacked gender information (3235 par-
wise meta-analysis (Pairwise MA). Furthermore, we ticipants). In addition, two multicentre studies
modelled all of the classes’ differences (across all of the stud- (Al-Duliamy et al., 2016; Lavelle, 1972) included samples
ies) using a single multivariate normal distribution with a from two and three different countries and from different
vector as mean and with a covariance matrix. After sampling continents and, consequently, they were counted as three
from the posterior distribution, the function returned fit sta- samples; however, in Lavelle (1972), the author did not
tistics that included adjusted estimates and associated credi- specify the African country, preventing it from being ana-
bility intervals (CrI; 2.5–97.5 percentiles) for normal lysed in meta-regression.
occlusion (adjusted), for all mean values of malocclusion Table 1 and Figure 2 show that only one study with nor-
classes and for malocclusion classes mean differences to normal occlusion participants (Lavelle, 1972) was published in
mal occlusion. We performed both random effects and fixed the 1970s, and thereafter was a lack of published reports for
effects Bayesian approaches. Both models fit well with coin- almost 30 years. In addition, the first author that investigated
cident point estimates and, recently, we selected the fixed- the Bolton ratios in Angle’s malocclusion groups was Crosby
effects variant results because this had the lowest deviance and Alexander (1989), more than 30 years after Bolton’s arti-
information criteria (DIC). Once again, I2 and Tau2 were cle (Bolton, 1958). After 1999, larger datasets were pub-
measured through random effects meta-analysis to account lished on different continents. Fifty-three cross-sectional
for the heterogeneity at each malocclusion subset. studies from four different continents, namely Africa, Asia,
The GRADE (Grading of Recommendations Assessment, South America and Europe were included in the qualitative
Development, and Evaluation) approach rated the quality of synthesis (Table 1). The studies by Lavelle (1972) and
evidence (Guyatt et al., 2011). The quality of evidence was Al-Duliamy et al. (2016) were multicentric studies that com-
based on five factors: (1) limitations of the study design or prised European, Asian and African participants, and Asian
the potential risk of bias across all studies; (2) inconsistency and African participants, respectively. Notably, no study was
of results (or heterogeneity); (3) indirectness (generalisabil- performed in Oceania or North America. Due to the inade-
ity); (4) imprecision (sufficient data); and (5) the potential quacy of continent representation, continent subgroup analy-
for publication bias (Schüneman et al., 2013). As recom- sis was not conceivable to perform.
mended by GRADE approach to systematic reviews, quality
score was not defined (Schüneman et al., 2013). Risk of bias. The Quality Assessment Tool for Observa-
tional Cohort and Cross-Sectional Studies statement pro-
posed by the NIHLBI score in the final sample of articles
Results was in the range of 6–8 out of 10 (as shown in Supplement
S4). Among the included studies, no study was of high
Study selection quality. Moreover, 47 articles presented medium quality, of
The initial electronic database search resulted in a total of which 11 articles presented 8 points (Asma, 2013;
2935 articles, leaving 2926 articles after the removal of Bugaighis et al., 2015; Cançado et al., 2015; Celikoglu
duplicates. No additional relevant articles were identified et al., 2013; Ismail and Abuaffan, 2015; Machado et al.,
after a hand search of reference lists. After title and abstract 2018; McSwiney et al., 2014; Nie and Lin, 1999; Shahid
screening, 190 studies were selected for full-text evalua- et al., 2016; Saritha et al., 2017; Ta et al., 2001) and 39
tion. After full-text eligibility assessment, 127 studies were articles presented 7 points (Al-Duliamy et al., 2016; Al
excluded (Supplement S3). Five studies were excluded for Sulaimani and Afify, 2006; Alkofide and Hashim, 2002;
presenting with low quality and high risk of bias (see the Carreiro et al., 2005; Chugh et al., 2015; Cİğer et al., 2006;
‘Risk of Bias’ in Materials and Methods section). Only six Crosby and Alexander, 1989; Elsheikhi et al., 2017; Endo
articles reported data for one Angle’s malocclusion, pre- et al., 2007, 2009, 2010; Fernandes et al., 2010; Freire
venting its inclusion in the Bayesian NMA. Finally, 53 et al., 2007; Hashim et al., 2015; Hyder et al., 2012; Jindal
studies were included in this review (Figure 1). and Bunger, 2013; Jóias and Scanavini, 2011; Jóias et al.,
Machado et al. 11
Figure 1. Flow chart of the study selection process.
2010; Kansal et al., 2012; Lavelle, 1972; Lee et al., 2011; 2013; Kansal et al., 2012; Lee et al., 2011; Machado et al.,
Lombardo et al., 2016; Machado et al., 2018; Manopa- 2018; Nie and Lin, 1999; Saritha et al., 2017; Ta et al.,
tanakul and Watanawirun, 2011; Maurya et al., 2015; Mol- 2001) and seven determined the sample size (Bugaighis
labashi et al., 2019; Mulimani et al., 2018; O’Mahony et al., et al., 2015; Cançado et al., 2015; Celikoglu et al., 2013;
2011; Oktay and Ulukaya, 2010; Patel et al., 2017; Ricci Ismail and Abuaffan, 2015; McSwiney et al., 2014; Shahid
et al., 2013; Sakoda et al., 2016; Shahid et al., 2016; Shastri et al., 2016; Ta et al., 2001). Strategies to minimise the
et al., 2015; Škrinjarić et al., 2018; Uysal et al., 2005; Vela potential sources of bias were not clearly described in most
et al., 2011; Zerouaoui et al., 2014). Five were of low qual- articles. Twelve articles fail to explain how they evaluated
ity (Anil and Monika, 2010; Devi et al., 2017; Kumar and intra- and/or inter-examiner errors or random error determi-
Chitra, 2017; Mirzakouchaki et al., 2007; Subbarao et al., nation (Anil and Monika, 2010; Devi et al., 2017; Kachoei
2014) and, consequently, were excluded. et al., 2011; Kansal et al., 2012; Kumar et al., 2015; Lee
Moreover, two articles were excluded (Kachoei et al., et al., 2011; Mirzakouchaki et al., 2007; Mollabashi et al.,
2011; Poosti and Jalali, 2007) due to abnormal SD values 2019; Mulimani et al., 2018; Poosti and Jalali, 2007;
(10 and 100 times lower than the mean value of SD pre- Škrinjarić et al., 2018; Subbarao et al., 2014).
sented in the remaining studies, respectively) (Supplements
S7–S10). These unusual SD values frame narrow confi-
dence intervals (CIs) gaining unreasonable weight in the
Meta-analysis findings
meta-analysis. We unsuccessfully tried to contact the Pairwise MA findings for normal occlusion. In normal occlu-
authors and, therefore, we decided to exclude these, pon- sion group, the assessment of OI and AI was sourced from
dering the likely negative consequences for the veracity of 25 and 26 studies, respectively (Figures 3 and 4). All the 26
the results. studies provided data for AI assessment, while one study
More specifically, only seven studies reported the (Jóias and Scanavini, 2011) had no data regarding OI.
setting, locations and relevant dates of cast models (Asma, Overall pooled results suggest an OI mean of 91.78% (95%
Table 1. Baseline characteristics for studies included in pairwise meta-analysis and Bayesian network meta-analysis.
12
Included in Included in
pairwise Bayesian
Participants Total OI Total AI meta- network meta-
Study (year) (country) (city) Continent (n) F/M Method (SD) (SD) F / M OI (SD) F / M AI (SD) analysis analysis
Normal occlusion
Mollabashi et al. (2019) Asia 60 30 / 30 Digital calliper 92.73 (1.74) 78.43 (2.31) NS NS Yes Yes
(Iran) (Hamadan) (0.01 mm)
Machado et al. (2018) Europe 29 10 / 19 Digital calliper 92.10 (2.20) 78.40 (3.50) 91.70 (2.20) / 92.9 (2.10) 77.90 (3.10) / 79.3 (4.10) Yes Yes
(Portugal) (Almada) (0.01 mm)
Patel et al. (2017) (India) Asia 50 25 / 25 Digital calliper 92.73 (2.69) 80.80 (2.86) 91.92 (3.40) / 93.55 (1.36) 80.34 (3.27) / 80.60 (2.45) Yes No
(Pune) (0.01 mm)
Sakoda et al. (2016) (Brazil) America 90 45 / 45 Digital calliper 91.63 (1.95) 77.57 (2.45) 91.35 (1.87) / 91.91 (2.04) 77.30 (2.28) / 77.85 (2.64) Yes No
(São Paulo) (0.01 mm)
Shahid et al. (2016) Asia 128 64 / 64 Digital calliper 92.80 (2.79) 79.25 (3.81) 93.10 (2.40) / 92.50 (3.10) 79.50 (3.60) / 79.00 (4.40) Yes No
(Pakistan) (different states) (0.01 mm)
Lombardo et al. (2016) Europe 56 22 / 34 Intraoral Scanner 91.56 (2.00) 77.65 (2.46) 91.55 (2.00) / 91.57 (2.00) 77.88 (2.00) / 77.30 (3.00) Yes No
(USA) (New York) (3shape)
Chugh et al. (2015) (India) Asia 50 25 / 25 Digital calliper 91.88 (1.99) 79.64 (2.61) 91.80 (2.34) / 91.96 (1.63) 79.16 (2.23) / 80.12 (1.73) Yes Yes
(Lucknow) (0.01 mm)
Bugaighis et al. (2015) (Libya) Africa 15 NS Digital calliper 90.24 (1.89) 76.88 (2.42) NS NS Yes Yes
(Benghazi) (0.01 mm)
Ismail and Abuaffan (2015) Africa 55 25 / 30 Digital calliper 91.47 (2.83) 77.46 (3.16) 91.25 (2.94) / 91.73 (2.90) 77.22 (3.43) / 77.73 (2.82) Yes Yes
(Sudan) (Khartoum) (0.1 mm)
Hashim et al. (2015) (Sudan) Africa 60 30 / 30 Digital calliper 90.80 (3.50) 76.90 (3.60) 90.60 (3.10) / 91.00 (3.90) 77.00 (3.70) / 76.90 (3.60) Yes No
(Khartoum) (0.01 mm)
Ricci et al. (2013) (Brazil) America 35 NS Digital calliper 90.38 (1.58) 77.49 (2.20) 90.36 (1.70) / 90.44 (1.20) 77.73 (2.39) / 76.68 (1.19) Yes Yes
Celikoglu et al. (2013) Europe 26 14 / 12 CBCT 90.69 (2.21) 77.58 (2.71) NS NS Yes No
(Turkey) (Karadeniz Ereğli)
Jóias and Scanavini (2011) America 35 8 / 27 Digital calliper NS 77.48 (2.22) NS 77.61 (2.45) / 77.05 (1.10) Yes No
(Brazil) (São Paulo) (0.01 mm)
Fernandes et al. (2011) America 140 70 / 70 Digital calliper 91.32 (1.98) 77.00 (2.71) 90.87 (1.94) / 91.77 (1.96) 76.54 (2.79) / 77.46 (2.61) Yes No
(Brazil) (Bauru) (0.01 mm)
Manopatanakul and Asia 37 NS Digital calliper 91.66 (1.74) 77.09 (2.18) NS NS Yes No
Watanawirun (2011) (0.01 mm)
(Thailand) (Bangkok)
Lee et al. (2011) (South Asia 307 188 / Digital calliper 90.42 (1.94) 77.54 (2.54) 90.30 (2.00) / 90.50 (1.90) 77.60 (2.60) / 77.50 (2.50) Yes No
Korea) (Seoul) 119 (0.01 mm)
Oktay and Ulukaya (2010) Europe 100 61 / 39 RMI 550 3D (0.01 92.10 (1.95) 79.28 (2.53) 91.63 (2.04) / 92.39 (1.84) 79.17 (2.65) / 79.35 (2.47) Yes Yes
(Turkey) (Erzurum) mm)
Journal of Orthodontics 47(1)
(Continued)
Table 1. (Continued)
pairwise Bayesian
Machado et al.
Jóias et al. (2010) (Brazil) America 35 8 / 27 Intraoral Scanner 91.58 (2.20) 78.66 (2.72) NS NS Yes No
(São Paulo) (3shape)
Freire et al. (2007) (Brazil) America 30 15 / 15 Digital calliper 91.46 (1.63) 77.83 (2.19) NS NS Yes No
(Rio de Janeiro) (0.01 mm)
Endo et al. (2007) (Japan) Asia 60 30 / 30 Digital calliper 91.60 (2.11) 78.39 (2.18) 91.69 (2.35) / 91.51 (1.88) 78.57 (2.19) / 78.21 (2.18) Yes No
(Niigata) (0.01 mm)
Cİğer et al. (2006) (Turkey) Europe 125 55 / 70 Digital calliper 91.95 (2.20) 77.95 (2.35) 91.82 (1.99) / 91.97 (1.65) 78.43 (2.41) / 78.62 (2.24) Yes Yes
(Hacettepe) (0.01 mm)
Carreiro et al. (2005) America 41 20 / 21 Microscribe 3DX 91.76 (2.51) 78.24 (3.40) NS NS Yes Yes
(Brazil) (Paraná)
Uysal et al. (2005) (Turkey) Europe 150 72 / 78 Digital calliper 91.90 (3.21) 78.56 (3.23) 91.73 (2.26) / 89.83 (2.33) 78.33 (2.42) / 78.18 (2.82) Yes Yes
(Konya) (0.01 mm)
Alkofide and Hashim (2002) Asia 60 NS Digital calliper 93.58 (2.12) 78.86 (2.55) 92.36 (2.37) / 92.12 (1.67) 78.79 (3.19) / 78.75 (2.27) Yes Yes
(Saudi Arabia) (Jeddah) (0.01 mm)
Nie and Lin (1999) (China) Asia 60 30 / 30 Software 93.27 (2.48) 81.52 (2.82) 93.11 (2.64) / 93.44 (2.35) 81.10 (2.27) / 81.95 (2.28) Yes Yes
(Beijing)
Lavelle (1972) Europe 40 20 / 20 Digital calliper 91.25 (2.00) 77.15 (1.60) 90.80 (1.85) / 91.70 (2.04) 77.50 (1.62) / 76.80 (1.49) Yes No
(0.1 mm)
Africa 40 20 / 20 Digital calliper 93.20 (2.11) 79.00 (2.02) 92.90 (1.78) / 93.50 (2.35) 78.60 (1.89) / 79.40 (2.06) Yes No
(0.1 mm)
Asia 40 20 / 20 Digital calliper 92.75 (1.53) 78.45 (1.55) 92.10 (1.55) / 92.60 (2.47) 78.20 (1.38) / 78.70 (1.66) Yes No
(0.1 mm)
Class I
Mollabashi et al. (2019) Asia 60 30 / 30 Digital calliper 92.61 (2.29) 78.79 (2.85) NS NS - Yes
Mulimani et al. (2018) Asia 15 NS Digital calliper 90.80 (2.15) 77.1 (2.3) NS NS - Yes
(Malaysia) (Melaka) (Indian) (0.01 mm)
Asia 27 NS Digital calliper 91.0 (1.72) 77.9 (2.3) NS NS - Yes
(Chinese) (0.01 mm)
(Malay) (0.01 mm)
Škrinjarić et al. (2018) Europe 39 20/19 ATOS II SO 91.57 (1.81) 78.19 (2.70) 91.24 (1.61) / 91.89 (1.99) 77.65 (2.49) / 78.75 (2.86) - Yes
(Croatia) (Zagreb)
Machado et al. (2018) Europe 50 29 / 21 Digital calliper 92.90 (2.70) 79.30 (4.00) 93.4 (2.30) / 92.50 (2.90) 79.60 (2.90) / 79.00 (4.60) - Yes
Saritha et al. (2017) (India) Asia 168 110 / 58 Digital calliper 92.38 (1.86) 79.37 (2.98) 92.39 (1.95) / 92.38 (1.82) 79.49 (2.37) / 79.30 (3.27) - Yes
(Telangana) (0.01 mm)
13
(Continued)
14
pairwise Bayesian
Mahmoud et al. (2017) Asia 52 NS Digital calliper 91.37 (2.98) 78.44 (2.91) NS NS - Yes
Elsheikhi et al. (2017) (Libya) Africa 20 10 / 10 Digital calliper 89.91 (1.79) 74.42 (2.06) NS NS - Yes
Cançado et al. (2016) (Brazil) America 321 NS Digital calliper 91.61 (2.04) 78.37 (2.68) NS NS - Yes
(Dourados) (0.01 mm)
Al-Duliamy et al. (2016) Asia 70 NS Digital calliper 91.23 (2.20) 78.72 (4.53) NS NS - Yes
(Iraq) (Baghdad) (0.01 mm)
Al-Duliamy et al. (2016) Africa 70 NS Digital calliper 91.63 (2.58) 78.85 (2.79) NS NS - Yes
(Egypt) (Cairo) (0.01 mm)
Chugh et al. (2015) (India) Asia 50 25 / 25 Digital calliper 93.06 (2.28) 79.60 (3.02) 93.35 (2.31) / 92.79 (2.28) 80.17 (3.13) / 79.09 (2.92) - Yes
(Lucknow) (0.01 mm)
Shastri et al. (2015) (India Asia 40 NS Digital calliper 91.73 (3.6) 76.89 (4.16) NS NS - Yes
(North)) (Lucknow)
Bughaighis et al. (2015) Africa 220 NS Digital calliper 91.55 (2.40) 78.29 (2.53) NS NS - Yes
(Libya) (Benghazi) (0.01 mm)
Ismail et al. (2015) (Sudan) Africa 49 26 / 23 Digital calliper 91.47 (2.83) 77.46 (3.16) 91.51 (3.27) / 91.39 (2.54) 77.00 (4.65) / 76.55 (3.34) - Yes
(Khartoum) (0.1 mm)
Maurya et al. (2015) (India) Asia 60 30 / 30 Digital calliper 92.38 (2.51) 80.13 (3.48) 93.03 (2.34) / 91.72 (2.58) 80.86 (3.28) / 79.40 (3.64) - Yes
(Madhya Pradesh) (0.01 mm)
Zerouaoui et al. (2014) Africa 30 NS Digital calliper 91.37 (2.05) 77.93 (2.60) NS NS - Yes
(Morocco) (Rabat)
Jindal and Bunger (2013) Asia 300 150 / Digital calliper 92.75 (3.15) 79.82 (3.85) 93.93 (3.34) / 91.58 (2.44) 80.87 (43135) / 78.77 - Yes
(India) (Punjab) 150 (0.01 mm) (3.38)
Asma (2013) (Malaysia) Asia 50 NS Digital calliper NS 78.83 (4.06) NS NS - Yes
(Selangor) (0.01 mm)
Ricci et al. (2013) (Brazil) America 35 NS Digital calliper 91.19 (2.70) 78.16 (2.87) 91.25 (3.24) / 91.17 (2.58) 78.66 (3.64) / 78.01 (2.66) - Yes
Hyder et al. (2012) Asia 40 20 / 20 Digital calliper 90.40 (2.69) 77.70 (2.81) 89.82 (3.06) / 91.06 (2.18) 77.92 (2.80) / 77.49 (2.87) - Yes
(Bangladesh) (Dhaka) (0.01 mm)
Kansal et al. (2012) (India) Asia 231 NS Digital calliper 91.80 (3.30) 79.20 (3.80) NS NS - Yes
(Karnataka) (0.01 mm)
O’Mahony et al. (2011) Europe 60 30 / 30 OrhoAnalyzer 92.30 (2.20) 79.00 92.4 (2.20) / 92.10 (2.20) 78.40 (2.90) / 79.60 (3.20) - Yes
(Ireland) (Cork) (Software) (43376)
Vela et al. (2011) (USA) America 207 110 / 97 Digital calliper NS 78.97 (2.29) NS 79.12 (1.99) / 78.84 (2.49) - No
(Texas) (0.01 mm)
Endo et al. (2010) (Japan) Asia 66 33 / 33 Digital calliper 91.10 (2.20) NS 91.18 (2.27) / 91.01 (2.17) NS - Yes
(Niigata) (0.01 mm)
(Continued)
pairwise Bayesian
Machado et al.
Oktay and Ulukaya (2010) Europe 100 65 / 35 RMI 550 3D (0.01 92.27 (2.16) 78.61 (2.80) 92.33 (1.88) / 92.24 (2.32) 78.66 (2.41) / 78.58 (3.01) - Yes
Strujić et al. (2009) (Croatia) Europe 110 68 / 42 Digital calliper 91.81 (1.99) 78.25 (2.58) NS NS - Yes
(Zagreb) (0.01 mm)
Endo et al. (2009) (Japan) Asia 101 59 / 42 Digital calliper 91.15 (2.14) 77.84 (2.46) 91.14 (2.33) / 91.15 (1.99) 77.97 (2.55) / 77.74 (2.39) - Yes
(Niigata) (0.01 mm)
(Niigata) (0.01 mm)
Al Sulaimani et al. (2006) Asia 98 62 / 36 Ortho-l software 93.90 (4.07) 81.11 (5.07) NS NS - Yes
(Saudi Arabia) (Jeddah)
Cİğer et al. (2006) (Turkey) Europe 125 70 / 55 Digital calliper 91.95 (2.20) 77.95 (2.35) 91.97 (1.65) / 91.82 (1.99) 78.62 (2.24) / 78.43 (2.41) - Yes
Carreiro et al. (2005) America 44 22 / 22 Microscribe 3DX 92.13 (2.08) 77.13 (3.15) NS NS - Yes
(Brazil) (Panamá)
Uysal et al. (2005) (Turkey) Europe 156 150 / 6 Digital calliper 91.90 (3.21) 78.56 (3.23) 91.65 (3.51) / 91.57 (2.98) 78.18 (3.31) / 78.44 (3.18) - Yes
(Konya) (0.01 mm)
Laino et al. (2003) (Italy) Europe 57 31 / 26 Digital calliper 91.72 (2.20) 78.12 (2.41) NS NS - Yes
(Campania) (0.01 mm)
Araújo et al. (2003) (Brazil) America 100 58 / 42 Digital calliper NS 78.18 (2.85) NS NS - Yes
(Belo Horizonte) (0.01 mm)
Alkofide and Hashim (2002) Asia 60 30 / 30 Digital calliper 92.24 (2.04) 78.77 (2.74) 92.12 (1.67) / 92.36 (2.37) 78.75 (2.27) / 78.79 (3.19) - Yes
Ta et al. (2001) (Hong Kong) Asia 50 25 / 25 Digital calliper 90.65 (1.19) 77.55 (1.80) 91.10 (1.00) / 90.20 (1.20) 77.60 (1.80) / 77.50 (1.80) - Yes
(Sheung Wan) (0.01 mm)
Nie and Lin (1999) (China) Asia 60 30 / 30 Software (0,01 93.27 (2.48) 81.52 (2.82) 93.62 (2.42) / 93.41 (2.53) 81.87 (2.51) / 81.25 (2.87) - Yes
(Beijing) mm)
Crosby and Alexander America 30 NS Digital calliper 91.30 (2.40) 77.2 (2.70) NS NS - Yes
(1989) (USA) (Texas) (0.01 mm)
Class II
Škrinjarić et al. (2018) Europe 55 34 / 23 ATOS II SO 91.27 (2.04) 77.46 (2.17) 91.50 (2.23) / 90.94 (1.71) 77.51 (1.93) / 77.39 (2.51) - Yes
(Croatia) (Zagreb)
(Chinese) (0.01 mm)
15
(Continued)
16
pairwise Bayesian

(Malay) (0.01 mm)
Saritha et al. (2017) (India) Asia 103 70 / 33 Digital calliper 92.296 78.642 92.14 (1.9) / 92.37 (2.05) 78.21 (2.56) / 78.84 (2.99) - Yes
(Telangana) (0.01 mm) (1.997) (2.868)
Cançado et al. (2016) (Brazil) America 324 NS Digital calliper 91.46 (2.06) 78.31 (2.39) NS NS - Yes
(Dourados) (0.01 mm)
Al-Duliamy et al. (2016) Asia 40 NS Digital calliper 91.54 (2.66) 79.05 (2.64) NS NS - Yes
(Iraq) (Baghdad) (0.01 mm)
Al-Duliamy et al. (2016) Africa 40 NS Digital calliper 89.14 (5.13) 78.46 (3.97) NS NS - Yes
(Egypt) (Cairo) (0.01 mm)
Shastri et al. (2015) (India Asia 50 NS Digital calliper 90.77 (2.13) 81.10 (5.01) NS NS - Yes
(North)) (Lucknow)
Ismail and Abuaffan (2015) Africa 59 27 / 22 Digital calliper 92.05 (3.11) 77.45 (4.8) 92.22 (3.84) / 91.92 (2.35) 77.17 (6.05) / 77.68 (3.44) - Yes
Maurya et al. (2015) (India) Asia 60 30 / 30 Digital calliper 91.69 (2.4) 79.06 (2.56) 91.67 (1.92) / 91.7 (2.87) 77.98 (1.95) / 80.14 (2.7) - Yes
(Madhya Pradesh) (0.01 mm)
Zerouaoui et al. (2014) Africa 30 NS Digital calliper 92.597 79.5975 NS NS - Yes
(Morocco) (Rabat) (2.41398) (2.94213)
McSwiney et al. (2014) Europe 60 30 / 30 Software (0.01 90.20 (2.27) 76.50 (2.77) 89.90 (2.31) / 90.50 (2.24) 76.50 (2.87) / 76.55 (2.70) - Yes
(Ireland) (Dublin) mm)
O’Mahony et al. (2011) Europe 120 60 / 60 OrhoAnalyzer 92.20 (2.19) 79.40 (3.36) 92.30 (2.31) / 92.05 (2.03) 79.50 (3.56) / 79.35 (3.20) - Yes
(Ireland) (Cork) (Software)
(Niigata) (0.01 mm)
(Continued)
pairwise Bayesian
Machado et al.
(Zagreb) (0.01 mm)
(Niigata) (0.01 mm)
(Niigata) (0.01 mm)
Al Sulaimani et al. (2006) Asia 52 34 / 18 Ortho-l software 93.06 (3.65) 81.88 (4.31) NS NS - Yes
Uysal et al. (2005) (Turkey) Europe 191 105 / 86 Digital calliper 91.27 (3.35) 78.59 (3.48) NS NS - Yes
(Konya) (0.01 mm)
Laino et al. (2004) (Italy) Europe 24 18 / 6 Digital calliper 91.24 (1.85) 78.04 (2.35) NS NS - Yes
Nie and Lin (1999) (China) Asia 120 60 / 60 Software (0.01 92.06 (2.50) 80.79 (3.19) 92.10 (2.66) / 92.02 (2.33) 80.69 (3.72) / 80.89 (2.54) - Yes
(Beijing) mm)
(1989) (USA) (Texas) (0.01 mm)
Class II – Division 1
Elsheikhi et al. (2017) (Libya) Africa 20 10 / 10 Digital calliper 91.38 (3.06) 76.29 (3.02) NS NS - Yes
Chugh et al. (2015) (India) Asia 40 20 / 20 Digital calliper 91.53 (2.49) 78.96 (3.56) 92.24 (2.43) / 90.83 (2.41) 79.95 (2.78) / 77.97 (3.66) - Yes
(Lucknow) (0.01 mm)
(Continued)
17
18
pairwise Bayesian

Ricci et al. (2013) (Brazil) America 35 NS Digital calliper 90.67 (2.4) 77.29 (2.51) 90.37 (2.35) / 90.76 (2.45) 77.27 (2.08) / 77.3 (2.65) - Yes
Cİğer et al. (2006) (Turkey) Europe 71 40 / 31 Digital calliper 90.83 (3.9) 78.04 (2.57) 90.54 (3.4) / 91.05 (4.24) 77.94 (2.46) / 78.11 (2.65) - Yes
Carreiro et al. (2005) America 54 26 / 28 Microscrrib 3DX 92.24 (2.56) 79.79 (4.24) NS NS - Yes
(Brazil) (Panamá)
Uysal et al. (2005) (Turkey) Europe 157 82 / 75 Digital calliper 91.12 (3.34) 78.50 (3.30) 91.19 (2.53) / 91.07 (3.96) 78.68 (3.06) / 78.33 (2.42) - Yes
(Konya) (0.01 mm)
(Beijing) mm)
(1989) (USA) (Texas) (0.01 mm)
Class II – Division 2
Machado et al. (2018) Europe 28 20 / 8 Software (0.01 91.60 (2.60) 78.60 (3.40) 92.3 (2.00) / 91.30 (2.90) 78.50 (3.70) / 78.70 (3.30) - Yes
(Portugal) (Almada) mm)
Mahmoud et al. (2017) Asia 3 NS Software (0.01 92.42 (2.17) 78.57 (1.53) NS NS - Yes
(Sudan) (Khartoum) mm)
Bughaighis et al. (2015) Africa 12 NS Software (0.01 91.56 (1.21) 78.20 (2.29) NS NS - Yes
(Libya) (Benghazi) mm)
Asma (2013) (Malaysia) Asia 50 NS Software (0.01 NS 80.33 (4.71) NS NS - Yes
(Selangor) mm)
Kansal et al. (2012) (India) Asia 17 NS Software (0.01 91.80 (2.90) 79.10 (3.00) NS NS - Yes
(Karnataka) mm)
(Continued)
pairwise Bayesian
Machado et al.
Uysal et al. (2005) (Turkey) Europe 34 23 / 11 Software (0.01 91.94 (3.34) 79 (4.23) 90.81 (2.27) / 89.81 (4.65) 79.63 (3.35) / 78.70 (4.64) - Yes
(Konya) mm)
(Beijing) mm)
Crosby and Alexander America 29 NS Software (0.01 91.50 (3.10) 76.80 (5.30) NS NS - Yes
(1989) (USA) (Texas) mm)
Class III
Škrinjarić (2018) (Croatia) Europe 27 15 /12 ATOS II SO 91.50 (2.88) 78.08 (3.03) 92.23 (2.50) / 90.92 (3.12) 78.62 (3.22) / 77.67 (2.10) - Yes
(Zagreb)
(Chinese) (0.01 mm)
(Malay) (0.01 mm)
Machado et al. (2018) Europe 38 25 / 13 Software (0.01 92.00 (2.00) 78.00 (2.90) 91.80 (1.60) / 92.10 (2.30) 78.00 (2.90) / 78.10 (2.90) - Yes
(Portugal) (Almada) mm)
Saritha et al. (2017) (India) Asia 40 21 / 19 Software (0.01 92.967 79.72 (2.52) 92.99 (1.75) / 92.94 (1.38) 79.92 (3.06) / 79.54 (1.97) - Yes
(Telangana) mm) (1.546)
Mahmoud et al. (2017) Asia 11 NS Software (0.01 91.38 (2.04) 78.37 (3.16) NS NS - Yes
(Sudan) (Khartoum) mm)
Elsheikhi et al. (2017) (Libya) Africa 20 10 / 10 Software (0.01 92.05 (2.96) 76.65 (4.09) NS NS - Yes
(Benghazi) mm)
Cançado et al. (2016) (Brazil) America 66 NS Software (0.01 91.22 (2.07) 77.90 (2.85) NS NS - Yes
(Dourados) mm)
Al-Duliamy et al. (2016) Asia 10 NS Software (0.01 91.82 (2.24) 78.80 (2.15) NS NS - Yes
(Iraq) (Baghdad) mm)
Al-Duliamy et al. (2016) Africa 10 NS Software (0.01 90.65 (3.71) 78.65 (4.20) NS NS - Yes
(Egypt) (Cairo) mm)
Chugh et al. (2015) (India) Asia 30 NS Software (0.01 94.05 (2.01) 81.23 (3.11) 94.48 (1.83) / 93.47 (2.11) 81.96 (3.17) / 80.49 (2.98) - Yes
(Lucknow) mm)
(Continued)
19
20
pairwise Bayesian
Shastri et al. (2015) (India Asia 20 NS Software (0.01 91.33 (2.32) 77.51 (5.64) NS NS - Yes
(North)) (Lucknow) mm)
Bughaighis et al. (2015) Africa 13 NS Software (0.01 90.97 (2.93) 77.48 (3.51) NS NS - Yes
(Libya) (Benghazi) mm)
Ismail et al. (2015) (Sudan) Africa 43 27 / 16 Software (0.01 92.60 (3.01) 77.71 (4.20) 93.58 (2.71) / 92.02 (3.03) 78.01 (4.12) / 77.53 (4.24) - Yes
(Khartoum) mm)
Maurya et al. (2015) (India) Asia 24 12 / 12 Software (0.01 94.72 (1.13) 84.49 (1.33) 95.51 (0.72) / 93.93 (0.88) 85.56 (0.93) / 84.33 (1.57) - Yes
(Madhya Pradesh) mm)
Zerouaoui et al. (2014) Africa 30 NS Software (0.01 92.075 78.2358 NS NS - Yes
(Morocco) (Rabat) mm) (2.2062) (2.85751)
McSwiney et al. (2014) Europe 60 NS Software (0.01 92.25 (2.31) 78.35 (2.42) 92.30 (2.27) / 92.20 (2.41) 78.15 (2.37) / 78.55 (2.58) - Yes
(Ireland) (Dublin) mm)
Asma (2013) (Malaysia) Asia 50 NS Software (0.01 NS 79.09 (2.82) NS NS - Yes
(Selangor) mm)
(Niigata) (0.01 mm)
(Zagreb) (0.01 mm)
(Niigata) (0.01 mm)
(Niigata) (0.01 mm)
Carreiro et al. (2005) America 46 23 / 23 Microscribe 3DX 92.30 (2.69) 79.54 (4.46) NS NS - Yes
(Brazil) (Panamá)
Uysal et al. (2005) (Turkey) Europe 113 55 / 58 Digital calliper 91.69 (3.66) 78.83 (3.46) 92.34 (3.67) / 91.01 (3,56) 79.59 (3.67) / 78.03 (3.06) - Yes
(Konya) (0.01 mm)
(Continued)
pairwise Bayesian
Machado et al.
Al Sulaimani et al. (2006) Asia 10 2/8 Ortho-l software 96.30 (1.45) 80.58 (3.74) NS NS - Yes
Laino et al. (2004) (Italy) Europe 13 6/7 Digital calliper 90.94 (2.26) 78.19 (2.27) NS NS - Yes
(Beijing) mm)
AI, Anterior Index; OI, Overall Index; CBCT, cone-beam computed tomography; NS, not stated; SD, standard deviation.
21
Figure 2. Datasets by year and population group by normal occlusion and Angle’s malocclusion groups. Area of the circle is
proportional to sample size.
Figure 3. Forest plot of studies with OI mean values for normal occlusion patients. Mean effect size estimates have been
calculated with 95% CIs and are shown in the figure. Area of squares represents sample size, continuous horizontal lines and
diamond width represents 95% CI. Blue diamond centre and the vertical red dotted line point to the overall pooled estimate.
Machado et al. 23
Figure 4. Forest plot of studies with AI mean values for normal occlusion patients. Mean effect size estimates have been
calculated with 95% CIs and are shown in the figure. Area of squares represents sample size, continuous horizontal lines and
diamond width represents 95% CI. Blue diamond centre and the vertical red dotted line point to the overall pooled estimate.
CI = 91.42–92.14) and an AI mean of 78.25% (95% CI = possible. Thus, we adopted a Bayesian NMA approach to
77.87–78.62). In both syntheses, heterogeneity was high (I2 pool all available direct and indirect comparisons between
= 92.87% and I2 = 90.67% in OI and AI, respectively). normal occlusion versus Angle’s Class I, Class II, Class II
Next, we looked for gender differences on OI and AI division 1, Class II division 2 and Class III values. The net-
through gender subgroup meta-analysis. Only Class I pre- work fit statistic outcome included adjusted normal occlu-
sented a gender impact with male patients having a higher sion mean values for each Angle’s malocclusions and the
OI (0.30, 95% CI = 0.00–0.59) and AI (0.41, 95% CI = estimated normal versus malocclusion difference of means,
0.00–0.83) mean values than female patients (Supplement with the degree of certainty of such differences reported as
S5). The GRADE assessment is depicted in Table 5. CrIs and heterogeneity as I2 and Tau2 (Tables 3 and 4).
The mean change from normal occlusion for AI means
Pairwise MA normal occlusion versus Bolton’s original val- Class I, Class II, Class II division 1 and Class II division 2
ues. Direct comparison of the Pairwise MA pooled esti- was −0.01 (95% CrI = −0.26 to 0.24), −0.10 (95% CrI =
mates for AI and OI normal occlusion mean values with −0.35 to 0.15), −0.14 (95% CrI = −0.42 to 0.14) and 0.19
Bolton’s original values, through Z-test, revealed signifi- (95% CrI = −0.17 to 0.55), respectively, but in all, the 95%
cant differences in both AI and OI (P < 0.10) (Table 2). CrI included zero (Table 3). A similar trend was also
observed for OI means when we compared the means of
Bayesian network meta-analysis findings for Angle’s malocclu- Class I, Class II and Class II division 2 with normal
sions groups. The difference in mean change for normal occlusion (Table 4).
occlusion (baseline) compared with different Angle’s mal- In contrast, we found a meaningful difference between
occlusion groups is presented in Tables 3 and 4, and can be Angle’s Class III versus normal occlusion for both OI and
seen as a measure of the average effort required to treat each AI means (0.76, 95% CrI = 0.55–0.98, and 0.61, 95% CrI
represented malocclusion towards a proportional occlusion. = 0.35–0.87, respectively) and for Class II division 1
In studies with no normal occlusion data, the comparison (−0.28, 95% CrI = −0.52 to −0.05) against normal occlu-
with the different types of Angle’s malocclusions was not sion AI means, since the null difference is not within the
Table 2. Comparison of Pairwise MA of normal occlusion with Bolton’s original values.
AI OI
n Mean (%) SD (%) P (Z-test) n Mean (%) SD (%) P (Z-test)
PMA normal occlusion 1954 78.25 8.5 < 0.001 1919 91.78 4.1 0.08
Bolton’s original values 55 77.20 1.65 55 91.3 1.91
Significant correlations identified in bold (P < 0.10).

AI, Anterior Index; MA, meta-analysis; OI, Overall Index; SD, standard deviation.
Table 3. Results of pairwise MA of normal occlusion and Bayesian NMA of Angle’s malocclusion groups and difference to normal
occlusion in the AI.
AI Heterogeneity
Mean difference to
normal occlusion (%) Studies
Model n Mean [95% CI] [95% CrI] (n) Tau2 [SE] I2
Random effects
Normal occlusion 1954 78.25 [77.87–78.62] - 28 0.896 [0.623] 90.67
Bayesian NMA
Normal occlusion 780 78.29 [61.64–95.05] - - - -
(network adjusted)
Class I 3425 78.26 [61.58–95.01] −0.01 [−0.26 to 0.24] 11 0.100 [0.165] 27.36
Class II 2717 78.10 [61.60–94.66] −0.10 [−0.35 to 0.15] 8 0.07 [0.160] 25.73
Class II / division 1 1111 78.07 [61.47–94.69] −0.14 [−0.42 to 0.14] 10 0.265 [0.265] 49.93
Class II / division 2 453 78.41 [61.85–95.08] 0.19 [−0.17 to 0.55] 6 0 [0.231] 0
Class III 1503 78.86 [62.14–95.51] 0.61 [0.35–0.87] 10 0.135 [0.207] 31.41
Bold values indicate significant difference to normal occlusion.

CI/CrI boundaries and I2 in %.
Malocclusion classes heterogeneity estimators were extracted from difference to normal Pairwise MA, while normal occlusion heterogeneity was
related to raw mean pairwise MA.
AI, Anterior Index; CI, confidence interval; CrI, credibility interval; MA, meta-analysis; NMA, network meta-analysis.
credibility region (Tables 3 and 4). The GRADE assess- years have been of increased interest (Figure 2). Our results
ment is described in Table 5. in patients with normal occlusion demonstrated that the OI
mean was 91.78% (95% CI = 91.42–92.14) and AI mean
was 78.25% (95% CI = 77.87–78.62) worldwide, while the
Additional analyses values proposed by Bolton (1958) were respectively
Funnel plots revealed no evidence of publication bias smaller.
(Supplement S6). Clinically, and according to Bolton’s original values, if
the OI are > 91.3% and > 77.2%, respectively, it indicates
mandibular tooth material excess. In this case, maxillary
Discussion teeth are relatively smaller compared to the mandibular
teeth, and mandibular teeth stripping or extraction can be
Summary of main findings pondered as a treatment option. Since Bolton’s original val-
To the best of our knowledge, this is the first systematic ues are not the same as those found in this meta-analysis for
review that attempted to estimate overall OI and AI values normal occlusion patients, then unnecessary stripping or
in patients with normal occlusion and Angle’s malocclu- teeth extraction can be performed, which can lead to clini-
sion. Despite the apparent gap in observational studies cal complications when trying to reach a stable occlusion
about normal occlusion between 1972 and 1998, the last 20 based on Andrew’s six keys.
Machado et al. 25
Table 4. Results of Pairwise MA of normal occlusion and Bayesian NMA of Angle’s malocclusion groups and difference to normal
occlusion in OI.
OI Heterogeneity
Mean difference to normal

Model n Mean [95% CI] occlusion (%) [95% CrI] Studies (n) Tau2 [SE] I2
Random effects
Normal occlusion 1919 91.78 [91.42–92.14] - 27 0.828 [0.712] 92.87
Bayesian NMA
Normal occlusion 780 91.78 [75.46–108.05] - - - -
(network adjusted)
Class I 3134 91.92 [75.72–108.10] 0.16 [−0.04 to 0.37] 11 0.455 [0.294] 72.24
Class II 2641 91.71 [75.37–108-08] −0.06 [−0.27 to 0.15] 8 0.412 [0.317] 73.39
Class II / division 1 1061 91.50 [75.35–107-65] −0.28 [−0.52 to −0.05] 10 0.356 [0.269] 65.09
Class II / division 2 403 91.70 [75.39–107.92] 0.01 [−0.28 to 0.30] 6 0.553 [0.513] 73.41
Class III 1419 92.52 [76.27–108.89] 0.76 [0.55–0.98] 10 0.887 [0.553] 82.63
Bold values indicate significant difference to normal occlusion.

Mean and CrIs boundaries in %.
Baseline normal occlusion is the covariate in the adjusted NMA.
Malocclusion classes heterogeneity estimators were extracted from difference to normal Pairwise MA, while normal occlusion heterogeneity was
related to raw mean pairwise MA.
CrI, credibility interval; MA, meta-analysis; NMA, network meta-analysis; OI, Overall Index.
Table 5. GRADE evidence profile for OI and AI in normal occlusion and malocclusions.
Outcome Study design Risk of bias Inconsistency Indirectness Imprecision Publication bias
Normal OI Observational Not serious Serious* Not serious Not serious Not serious
occlusion studies (serious)
AI Observational Not serious Serious* Not serious Not serious Not serious
studies (serious)
Malocclusions OI Observational Not serious Serious* Not serious Not serious Not serious
studies (serious)
AI Observational Not serious Serious* Not serious Not serious Not serious
studies (serious)
*Downgraded for serious inconsistency: even considering the large sample sizes and the use of digital, some degree of heterogeneity is still
perceptible.
AI, Anterior Index; GRADE, Grading of Recommendations Assessment, Development, and Evaluation; I Explanations: GRADE Working Group
grades of evidence; OI, Overall Index.
In fact, pooled Pairwise MA normal occlusion estimates representative of the continent as a whole. There is a lack of
were not significantly different from Bolton’s values; how- studies in the continents of North America and Oceania.
ever, this direct comparison, though necessary, is quite Similarly, the African continent is portrayed only by Libya,
unfair and disproportionate as shown by the extraordinarily Egypt, Morocco and Sudan, the European continent is repre-
low statistical power. Also, the computed SDs from meta- sented mainly by studies in Turkey and Ireland, the Asian
analytical pooled estimates revealed very discrepant and continent is mostly represented by studies from India, and the
elevated values when compared with Bolton’s, and we American continents only had two studies from North
believe that these direct comparisons are biased since it is America and the remaining studies are from Brazil. Despite
not adequate to compare so unequal samples. this restriction, future research should address race and
It is important to remark that we have not made geographic genetic backgrounds to weigh their influence on the mesio-
subgroups based on continent since the studies are not fully distal proportions since it was not possible to perform this in
this study due to the lack of such data. Still, globalisation and opinion, these results can be obtained from the lack of high
inter-racial/inter-ethnic mixing strongly support the concept methodological homogeneity, but also from a high varia-
of non-static proportions and the necessity for continued bleness of mesiodistal width proportions among the
research. Further, gender and geographic location, in general, populations.
are not factors that influence dental width proportions. Regarding methodology, most studies took teeth meas-
Regarding the relation between normal occlusion and urements from plaster models. Only one investigation used
Angle malocclusions, in general, our results determined that cone-beam computed tomography (CBCT) (Celikoglu
no significant difference in the tooth size discrepancy existed et al., 2013), five studies have digitised plaster models and
for the OI and AI between normal occlusion and different mal- subsequently performed the measurements (Al Sulaimani
occlusion groups, except for the Class III malocclusion in both and Afify, 2006; Jóias et al., 2010; Lombardo et al., 2016;
AI and OI, and Class I malocclusion in OI only. Under these O’Mahony et al., 2011; Škrinjarić et al., 2018) and three
circumstances, the results indicate that the discrepancy of also used an electronic measuring device (Carreiro et al.,
intermaxillary tooth size may be one of the important factors 2005; Nie and Lin, 1999; Oktay and Ulukaya, 2010).
in the cause of malocclusions, especially in Angle’s Class III. Although in the past calliper measurement in plaster mod-
The results also suggest that these OI and AI differences els was the gold standard, nowadays the study of models
for normal occlusion in Angle’s Class I and Class III may with virtual three-dimensional (3D) technology has higher
be explained by upper and/or lower discrepancy. For both reliability and accuracy (Aragón et al., 2016; De Luca
Angle’s Class I and III difference for normal occlusion, a Canto et al., 2015; Fleming et al., 2011; Luu et al., 2012)
possible clinical explanation for this discrepancy may be and should be used as the first choice for diagnosis and
due to smaller mesiodistal maxillary tooth sizes and/or treatment planning in orthodontics, specifically to deter-
greater mesiodistal mandibular widths. mine the width of teeth. Additionally, study models pro-
duced by CBCT are far from perfect for replacing digital
models. Hence, in the future, with proper improvement,
Quality of the evidence and potential biases CBCT will ensure a multiplicity of analyses from a single
in the review process record (Ferreira et al., 2017). Furthermore, it is imperative
All studies included in the meta-analysis presented overall that, in addition to the mesiodistal width, the labio-lingual
medium quality, according to our predefined quality assess- and inclination data should be evaluated since they may
ment and risk of bias. However, there are important matters also present great variability in populations. Thus, a 3D
that need to be pointed out. Hypothetical limitations would orthodontic diagnosis and treatment plan is more desirable
be the fact that this systematic review only contains obser- than a two-dimensional assessment.
vational studies and language biases, because only studies in The overall results of this study are in line with a previ-
English were included. However, except for restorative or ous comprehensive review (Othman and Harradine, 2006).
traumatic reasons, mesiodistal width of teeth remains pro- Although gender has no clinically significant effect on
spectively unchanged. Therefore, randomised controlled tri- tooth size discrepancy (TSD) in general, this study con-
als, and prospective or retrospective studies on this theme, firmed a significant difference in Class I malocclusion
would be inappropriate unless they were the result of a sec- between men and women. Besides, we proved what was
ondary observation. comprehensively stated that Class III malocclusions have
On the other hand, we must emphasise that most stud- higher average ratios (Othman and Harradine, 2006).
ies lack sample size calculation and are non-representa-
tive of the population, but are rather from an academic
setting. Besides, too many studies show a lack of informa- Conclusions
tion on calibration method or the number of examiners. The results of this systematic review show that overall
These items are extremely important to minimise selec- pooled OI and AI mean values for normal occlusion patients
tion bias and strengthen the generalisation of results, and are slightly above Bolton’s original values. Class I, for OI
its absence weakens the results of this systematic review. mean values, and Class III, for both OI and AI, are propor-
Additionally, no study has reported the existence of blind- tionally larger than normal occlusion patients. Gender had
ing examiners, since presumably the researchers them- almost no impact on teeth mesiodistal proportion.
selves were involved in measurements of teeth and
Angle’s evaluation. This potential bias should be consid-
ered in future research. Implications for clinical practice and
Significantly, the heterogeneity revealed by our meta- research
analysis refers, conceptually, to the variation in study out- Despite being one of several measures used in orthodon-
comes between studies. This variation could flag some tic planning, the results of this systematic review suggest
problems; however, we need to carefully assess this dis- that Bolton’s original values may be slightly underesti-
crepancy, contrary to common meta-analysis. In our mated as the OI and AI global standard. The use of
Machado et al. 27
inadequate standard measures for the dental proportion of Data availability

each population can lead to diagnostic errors and could Data are available at https://doi.org/10.5281/zenodo.3407853
influence the patient’s treatment outcome. Also, despite
these AI and OI mean values being originally developed Supplemental material
only for tooth width reduction (through interproximal
Supplemental material for this article is available online.
stripping or extraction), several patients with mesiodistal
disproportionality, mainly due to microdontia or agenesis
References
in one or more teeth, require post-orthodontic rehabilita-
Al-Duliamy MJ, Othman SS and Hussien FA (2016) Comparison of
tion treatments. Thus, in future, it is imperative to estab-
Bolton’s ratios in a sample of Iraqi and Egyptian populations. Journal
lish normative data for different malocclusions and their of Baghdad College of Dentistry 28: 172–175.
impact on proportion management during orthodontic Al-Khateeb S and Abu Alhaija E (2006) Tooth size discrepancies and arch
treatment. parameters among different malocclusions in a Jordanian sample.
As a result, in the future, there is a clear need for further Angle Orthodontist 76: 459–465.
Al-Omari I, Al-Bitar Z and Hamdan A (2008) Tooth size discrepancies
studies with more stringent methodologies with regard to
among Jordanian schoolchildren. European Journal of Orthodontics
sample size calculation, more representative population 30: 527–531.
samples, explicit calibration methods to reduce risk bias, Al Sulaimani FFH and Afify AR (2006) Bolton Analysis in Different
fostering the use of digital systems, and greater focus on the Classes of Malocclusion in a Saudi Arabian Sample. Egyption Dental
race of the population being studied combined with genetic Journal 52: 1119–1125.
Alkofide E and Hashim H (2002) Intermaxillary tooth size discrepancies
background analysis of the patients. In addition, the ques-
among different malocclusion classes: a comparative study. Journal
tion arises of the importance of orthodontic consensus to of Clinical Pediatric Dentistry 26: 383–387.
those who must establish normative data, since the results Altman DG, Machin D, Bryant T and Gardner M (eds) (2000) Statistics
of this study point to a difference from the original Bolton’s with Confidence. 2nd edn. London: BMJ Books.
values. Still, the evolution of our species by the inter-racial/ Anil S and Monika M (2010) Bolton Analysis of Himachali Ethnic
Population. Indian Journal of Dental Sciences 2: 12–14.
inter-ethnic mixing due to globalisation makes pressing the
Aragón MLC, Pontes LF, Bichara LM, Flores-Mir C and Normando D
need for continuous research on human proportions and, in (2016) Validity and reliability of intraoral scanners compared to
this case, on the mesiodistal teeth proportion, since, appar- conventional gypsum models measurements: A systematic review.
ently, these are not immutable. European Journal of Orthodontics 38: 429–434.
In forthcoming investigations, we believe it will be sub- Araujo E and Souki M (2003) Bolton anterior tooth size discrepancies
among different malocclusion groups. The Angle Orthodontist 73:
stantial to investigate the pre-orthodontic patient, in order to
307–313.
seek the ideal post-orthodontic position of each tooth. In this Asma A (2013) Comparison of anterior tooth size discrepancies among
way, we will know which tooth or teeth need mesiodistal different malocclusion groups. Malaysian Journal of Medicine and
intervention so that we can achieve normal occlusion with Health Sciences 9: 73–79.
proper mesiodistal proportion respecting Andrew’s six keys. Barbato L, Kalemai Z, Buti J, Baccini M, La Marca M, Duvina M, et al.
(2015) Effect of surgical intervention for removal of mandibular third
molar on periodontal healing of adjacent mandibular second molar :
Acknowledgements a systematic review and Bayesian network meta-analysis. Journal of
The authors thank Behnam Mirzakouchaki (Tabriz Medical Periodontology 87: 291–302.
Science University, Iran), Iman Bughaighis (Benghazi University, Beller EM, Glasziou PP, Altman DG, Hopewell S, Bastian H, Chalmers I,
Libya), Karine Sakoda (University of São Paulo, Brazil), Pavan et al. (2013) PRISMA for Abstracts: reporting systematic reviews in
journal and conference abstracts. PLoS Medicine 10: e1001419.
Kumar (Army College of Dental Sciences, India), Vinay Chugh
Bernabé E, Major PW and Flores-Mir C (2004) Tooth-width ratio dis-
(University of Connecticut Health Center, USA) and Vivekanand
crepancies in a sample of Peruvian adolescents. American Journal of
Kattimani (Institute of Dental Sciences, India) for providing Orthodontics and Dentofacial Orthopedics 125: 361–365.
additional data/clarifications on their studies and additional
Bishara SE, Jakobsen JR, Abdallah EM and Fernandez Garcia A (1989)
outcomes not reported in the paper. Comparisons of mesiodistal and buccolingnal crown dimensions of
the permanent teeth in three populations from Egypt, Mexico, and
Declaration of conflicting interests the United States. American Journal of Orthodontics and Dentofacial
Orthopedics 96: 416–422.
The author(s) declared no potential conflicts of interest with Black G (1902) Descriptive anatomy of human teeth. 4th edn.
respect to the research, authorship, and/or publication of this Philadelphia, PA: Wilmington Dental Manufacturing Co.
article. Bolton W (1958) Disharmony in tooth size and its relation to the analysis
and treatment of malocclusion. The Angle Orthodontist 28: 113–130.
Funding Bolton W (1962) The clinical application of a tooth-size analysis.
American Journal of Orthodontics 48: 504–529.
The author(s) received no financial support for the research, Bonwill W (1899) The scientific articulation of the human teeth as
authorship, and/or publication of this article. founded on geometrical, mathematical and mechanical laws. Dental
Items of Interest 1: 617–643.
Bugaighis I, Karanth D and Borzabadi-Farahani A (2015) Tooth size
ORCID iD discrepancy in a Libyan population, a cross-sectional study in school-
Vanessa Machado https://orcid.org/0000-0003-2503-260X children. Journal of Clinical and Experimental Dentistry 7: e100–e105.
Cançado RH, Goncalves Junior W, Valarelli FP, Freitas KM and Crespo JA Freire SM, Nishio C, Mendes Ade M, Quintao CC and Almeida MA
(2015) ‘Association between Bolton discrepancy and Angle maloc- (2007) Relationship between dental size and normal occlusion in
clusions. Brazilian Oral Research 29: 1–6. Brazilian patients. Brazilian Dental Journal 18: 253–257.
Carreiro LS, Santos-Pinto A, Raveli DB and Martins LP (2005) A discrep- Furuya-Kanamori L, Barendregt JJ and Doi SAR (2018) A new improved
ância de tamanho dentário, de Bolton, na oclusão normal e nos difer- graphical and quantitative method for detecting bias in meta-analysis.
entes tipos de más oclusões, bem como sua relação com a forma de International Journal of Evidence-Based Healthcare 16: 195–203.
arco e o posicionamento dentário. Revista Dental Press de Ortodontia Gilpatric WH (1923) Arch Predetermination—Is it Practical? The Journal
e Ortopedia Facial 10: 97–117. of the American Dental Association 10: 553–572.
Celikoglu M, Nur M, Kilkis D, Sezgin OS and Bayram M (2013) Mesiodistal Guyatt G, Oxman AD, Aki EA, Kunz R, Vist G, Brozek J, et al. (2011)
tooth dimensions and anterior and overall Bolton ratios evaluated by cone GRADE guidelines: 1. IntroductiondGRADE evidence profiles and
beam computed tomography. Australian Orthodontic Journal 29: 11–16. summary of findings tables. Journal of Clinical Epidemiology 64:
Chugh V, Chugh A, Maurya RP, Nagar A, Sharma V and Tandon P (2015) 383–394.
An evaluation of tooth size discrepancies among different malocclusion Hashim AA, Eldin A-H and Hashim H (2015) Bolton tooth size ratio
groups in North Indians. Journal of Orthodontic Research 3: 119–123. among Sudanese Population sample: A preliminary study. Journal of
Cİğer S, Aksu M and SaĞlam B (2006) Interarch Tooth-Size Relationships Orthodontic Science 4: 77–82.
of Normal Occlusion and Class II Division 1 Malocclusion Patients Hattab FN, Al-Khateeb S and Sultan I (1996) Mesiodistal crown diameters
in a Turkish Population Türk Populasyonunda Sınıf II Divizyon 1 of permanent teeth in Jordanians. Archives of Oral Biology 41: 641–645.
Maloklüzyon ve Normal Okluzyonlu Bireylerde İnterark Diş Boyutu Higgins J and Green S (2011) Cochrane Handbook for Systematic Reviews
İlişkileri’, Hacettepe Dişhekimliği Fakültesi Dergisi 30: 25–32. of Interventions. 5.1.0. Oxford: Cochrane Group.
Crosby DR and Alexander CG (1989) The occurrence of tooth size dis- Higgins JP, Thompson SG, Deeks JJ and Altman DG (2003) Measuring
crepancies among different malocclusion groups. American Journal inconsistency in meta-analyses. BMJ 327: 557–560.
of Orthodontics and Dentofacial Orthopedics 95: 457–461. Hughes T, Dempsey P, Richards L and Townsend G (2000) Genetic anal-
De Luca Canto G, Pacheco-Pereira C, Lagravere MO, Flores-Mir C and ysis of deciduous tooth size in Australian twins. Archives of Oral
Major PW (2015) Intra-arch dimensional measurement validity of laser- Biology 45: 997–1004.
scanned digital dental models compared with the original plaster models: Hyder M, Mamun M and Hossain M (2012) Tooth Size Discrepancies
A systematic review. Orthodontics and Craniofacial Research 18: 65–76. among Different Malocclusions in a Bangladeshi Orthodontic
Dempsey P and Townsend GC (2001) Genetic and environmental contri- population. Bangladesh Journal of Orthodontics and Dentofacial
butions to variation in human tooth size. Heredity 86: 685–693. Orthopedics 2: 8–17.
Devi LB, Singh OB and Laiphangbam J (2017) Evaluation of anterior and Ismail MA and Abuaffan AH (2015) Tooth size discrepancy among dif-
overall tooth size ratios by Bolton’s analysis in North Eastern Indian pop- ferent malocclusion groups in Sudanese sample. Orthodontic Waves
ulation. Journal of Pierre Fauchard Academy (India Section) 31: 47–50. 74: 37–41.
Doi SA, Barendreqt JJ, Khan S, Thalib L and Williams GM (2015a) Jindal R and Bunger E (2013) Bolton’s intermaxillary tooth size ratios
Advances in the meta-analysis of heterogeneous clinical trials I: The among school going children in Punjab population. Indian Journal of
inverse variance heterogeneity model. Contemporary Clinical Trials Oral Sciences 4: 110–113.
45: 130–138. Jóias RP and Scanavini MA (2011) Factors related to Bolton’s anterior
Doi SA, Barendreqt JJ, Khan S, Thalib L and Williams GM (2015b) ratio in Brazilians with natural normal occlusion. Brazilian Journal of
Advances in the meta-analysis of heterogeneous clinical trials II: The Oral Sciences 10: 69–73.
quality effects model. Contemporary Clinical Trials 45: 123–129. Jóias RP, Velasco LG, Scanavini MA, de Miranda AL and Siquiera DF (2010)
Egger M and Smith GD (1998) Bias in location and selection of studies. Evaluation of the Bolton ratios on 3D dental casts of Brazilians with
BMJ 316: 61–66. natural, normal occlusions. World Journal of Orthodontics 11: 67–70.
Elsheikhi F, Bugaighis I and Hamed T (2017) Tooth size discrepancy Kachoei M, Ahangar-Atashi MH and Pourkhamneh S (2011) Bolton’s
in a different malocclusion groups in Libya: a pilot study. Libyan intermaxillary tooth size ratios among Iranian schoolchildren.
International Medical University Journal 2: 92–100. Medicina Oral, Patologia Oral y Cirugia Bucal 16: 568–572.
Endo T, Abe R, Kuroki H, Oka K and Shimooka S (2008) Tooth size dis- Kansal A, Kittur N and Keluskar KM (2012) Analysis of Bolton’s ratio
crepancies among different malocclusions in a Japanese orthodontic among different malocclusion groups: A hospital based study. Indian
population. Angle Orthodontist 78: 994–999. Journal of Dentistry 3: 139–144.
Endo T, Ishida K, Shundo I, Sakaeda K and Shimooka S (2010) Effects Kesling H (1945) The philosophy of the tooth positioning appliance.
of premolar extractions on Bolton overall ratios and tooth-size dis- American Journal of Orthodontics and Oral Surgery 31: 297–304.
crepancies in a Japanese orthodontic population. American Journal of Kotsakis GA, Lian Q, Ioannou AL, Michalowicz BS, John MT and Chu H
Orthodontics and Dentofacial Orthopedics 137: 508–514. (2018) A network meta-analysis of interproximal oral hygiene meth-
Endo T, Shundo I, Abe R, Ishida K, Yoshino S and Shimooka S (2007) ods in the reduction of clinical indices of inflammation. Journal of
Applicability of Bolton’s tooth size ratios to a Japanese orthodontic Periodontology 89: 558–570.
population. Odontology 95: 57–60. Kravitz ND, Kusnoto B, BeGole E, Obrez A and Agran B (2009) How
Endo T, Uchikura K, Ishida K, Shundo I, Sakaeda K and Shimooka S well does Invisalign work? A prospective clinical study evaluating
(2009) Thresholds for clinically significant tooth-size discrepancy. the efficacy of tooth movement with Invisalign. American Journal of
Angle Orthodontist 79: 740–746. Orthodontics and Dentofacial Orthopedics 135: 27–35.
Fernandes TM, Janson G, Pinzan A, Sathler R, de Freitas LM and de Kumar MS, Mohan AM, Kommi PB, Venkatesan R, Suresh V and Arun
Freitas MR (2010) Applicability of Bolton tooth-size ratios in racial Kumar K (2015) Evaluation of Bolton’s Discrepancy in Un-treated
groups. World Journal of Orthodontics 11: e57–e62. Angles Class I Patients in Pondicherry population : A Cross-Sectional
Ferreira JB, Christovam IO, Alencar DS, da Motta AFJ, Mattos CT and Study. Journal of International Oral Health 7: 86–89.
Cury-Saramago A (2017) Accuracy and reproducibility of dental Kumar P and Chitra P (2017) Determination of Bolton Norms for Indian
measurements on tomographic digital models: A systematic review Population Sample. Orthodontic Journal of Nepal 7: 33–36.
and meta-analysis. Dentomaxillofacial Radiology 46: 20160455. Kumar P, Singh V, Kumar P, Sharma P and Sharma R (2013) Effects
Fleming PS, Marinho V and Johal A (2011) Orthodontic measurements of premolar extractions on Bolton overall ratios and tooth-size
on digital study models compared with plaster models: A systematic discrepancies in a Japanese orthodontic population. Journal of

review. Orthodontics and Craniofacial Research 14: 1–16. Orthodontic Science 2: 23–27.
Machado et al. 29
Laino A, Quaremba G, Paduano S and Stanzione S. (2003). Prevalence Patel Y, Nair V and Jamenis S (2017) Bolton analysis of the maratha popu-
of tooth-size discrepancy among different malocclusion groups. lation in Pune. Journal of Dental and Allied Sciences 6: 8–11.
Progress in Orthodontics 4: 37–44. Poosti M and Jalali T (2007) Tooth size and arch dimension in uncrowded
Lavelle C (1972) Maxillary and mandibular tooth size in different racial versus crowded class I malocclusions. The Journal of Contemporary
groups and in different occlusal categories. American Journal of Dental Practice 8(3): 1–8.
Orthodontics 61: 29–37. Ricci ID, Scanavini MA, Kaieda AK, Rosario HD and Paranhos LR
Lee SJ, Ahn SJ, Lim WH, Lee S, Lim J and Park HJ (2011) Variation of the (2013) Bolton ratio in subjects with normal occlusion and malocclu-
intermaxillary tooth-size relationship in normal occlusion. European sion. Brazilian Journal of Oral Sciences 12: 357–361.
Journal of Orthodontics 33: 9–14. Sakoda KL, Pinzan A, Janson G and Cury SEN (2016) Applicability of
Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, Bolton’s tooth size ratios in Mediterranean, Japanese and Japanese-
et al. (2009) The PRISMA statement for reporting systematic reviews Brazilian populations. Brazilian Journal of Oral Sciences 15: 269–272.
and meta-analyses of studies that evaluate health care interventions: Santoro M, Ayoub ME, Pardi VA and Cangialosi TJ (2000) Mesiodistal
Explanation and elaboration. PLoS Med 6: e1000100. crown dimensions and tooth size discrepancy of the permanent denti-
Lombardo L, Arreghini A, Ramina F, Huanca Ghislanzoni LT and Siciliani tion of Dominican Americans. Angle Orthodontist 70: 303–307.
G (2017) Predictability of orthodontic movement with orthodontic Saritha T, Sunitha C, Kiran Kumar P and Naveen R (2017) Applicability
aligners: a retrospective study. Progress in Orthodontics 18: 35. of Bolton’s Analysis to a South Telangana Population. Indian Journal
Lombardo L, Marcon M, Arveda N, La Falce G, Tonello E and Siciliani G of Dental Sciences 9: 225–232.
(2016) Preliminary biometric analysis of mesiodistal tooth dimensions Scheid RC and Weiss G (2012) Woelfel’s Dental Anatomy. 8th edn.
in subjects with normal occlusion. American Journal of Orthodontics Philadelphia, PA: Lippincott, Williams & Wilkins.
and Dentofacial Orthopedics 150: 105–115. Shahid F, Alam MK and Khamis MF (2016) Intermaxillary tooth size dis-
Luu NS, Nikolcheva LG, Retrouvey JM, Flores-Mir C, El-Bialy T, Carey crepancy in a Pakistani population: A stereomicroscope versus digital
JP, et al. (2012) Linear measurements using virtual study models. A caliper. European Journal of Dentistry 10: 176–182.
systematic review. Angle Orthodontist 82: 1098–1106. Shastri D, Singh A and Tandon P (2015) Bolton ratio in a North Indian
Machado V, Botelho J, Pereira D, Vasques M, Fernandes-Retto P, Proenca population with different malocclusions. Journal of Orthodontic
L, et al. (2018) Bolton ratios in Portuguese subjects among different Science 4: 83–85.
malocclusion groups. Journal of Clinical and Experimental Dentistry Škrinjarić A, Šlaj M and Šlaj M (2018) Tooth Size Discrepancies
10: e864–e868. and Dental Asymmetry in Different Malocclusions. Collegium
Mahmoud N, Eltahir H and Mageet A (2017) Tooth size discrepancy Antropologicum 42: 45–51.
among different malocclusion groups in Sudanese sample. Journal of Smith S, Buschang P and Watanabe E (2000) Interarch tooth size rela-
Orthodontics & Endodontics 3: 10. tionships of 3 populations: “does Bolton’s analysis apply”? American
Manopatanakul S and Watanawirun N (2011) Comprehensive intermax- Journal of Orthodontics and Dentofacial Orthopedics 117: 169–174.
illary tooth width proportion of Bangkok residents. Brazilian Oral Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, et al. (2011)
Research 25: 122–127. Recommendations for examining and interpreting funnel plot asym-
Maurya R, Gupta A, Garg J and Mishra H (2015) Seventh key of occlu- metry in meta-analyses of randomised controlled trials. BMJ 343:
sion: Diagnostic significance in different angle′s class I, II and III d4002.
malocclusions. Journal of Orthodontic Research 3: 188–191. Strujić M, Anić-Milošević S, Meštrović S and Šlaj M (2009) Tooth size
McSwiney TP, Millett DT, McIntyre GT, Barry MK and Cronin MS discrepancy in orthodontic patients among different malocclusion
(2014) Tooth size discrepancies in Class II division 1 and Class III groups. The European Journal of Orthodontics 31: 584–589.
malocclusion requiring surgical-orthodontic or orthodontic treatment. Subbarao VV, Regalla RR, Santi V, Anita G and Kattimani VS (2014)
Journal of Orthodontics 41: 118–123. Interarch tooth size relationship of indian population: Does Bolton’s anal-
Mirzakouchaki B, Shahrbaf S and Talebiyan R (2007) Determining tooth ysis apply? The Journal of Contemporary Dental Practice 15: 103–107.
size ratio in an Iranian-Azari population. Journal of Contemporary Schünemann H, Brożek J, Guyatt G and Oxman A (eds) GRADE hand-
Dental Practice 8: 86–93. book for grading quality of evidence and strength of recommenda-
Mollabashi V, Soltani MK, Moslemian N, Akhlaghian M, Akbarzadeh tions. Updated October 2013.
M, Samavat H, et al. (2019) Comparison of Bolton ratio in normal Ta TA, Ling JYK and Hägg U (2001) Tooth-size discrepancies among
occlusion and different malocclusion groups in Iranian population. different occlusion groups of southern Chinese children. American
International Orthodontics 17: 143–150. Journal of Orthodontics and Dentofacial Orthopedics 120: 556–558.
Mulimani PS, Azmi M, Jamali N, Basir N and Soe H (2018) Bolton’s tooth Tedesco TK, Giminez T, Floriano I, Montagner AF, Camargo LB, Calvo
size discrepancy in Malaysian orthodontic patients: are occlusal char- AFB, et al. (2018) Scientific evidence for the management of dentin
acteristics such as overjet, overbite, midline, and crowding related caries lesions in pediatric dentistry : A systematic review and network
to tooth size discrepancy in specific malocclusions and ethnicities? meta-analysis. PLoS ONE 13: e0206296.
Trends in Orthodontics 8: 36–43. Uysal T, Sari Z, Basciftci FA and Memili B (2005) Intermaxillary
Neff CW (1949) Tailored occlusion with the anterior coefficient. American tooth size discrepancy and malocclusion: Is there a relation? Angle
Journal of Orthodontics 35: 309–313. Orthodontist 75: 208–213.
Neff CW (1957) The size relationship between the maxillary and mandibular Vela E, Taylor RW, Campbell PM and Buschang PH (2011) Differences
anterior segments of the dental arch. Angle Orthodontist 27: 138–147. in craniofacial and dental characteristics of adolescent Mexican
Nie Q and Lin J (1999) Comparison of anterior tooth size discrepan- Americans and European Americans. American Journal of
cies among different malocclusion groups. American Journal of Orthodontics and Dentofacial Orthopedics 140: 839–847.
Orthodontics and Dentofacial Orthopedics 116: 539–544. Wallace BC, Lajeunesse MJ, Dietz G, Dahabreh IJ, Trikalinos TA, Schmid
O’Mahony G, Millett DT, Barry MK, McIntyre GT and Cronin MS (2011) CH, et al. (2017) OpenMEE: Intuitive, open-source software for meta-
Tooth size discrepancies in Irish orthodontic patients among different analysis in ecology and evolutionary biology. Methods in Ecology and
malocclusion groups. Angle Orthodontist 81: 130–133. Evolution 8: 941–947.
Oktay H and Ulukaya E (2010) Intermaxillary tooth size discrepan- Young J (1923) Rational treatment of infraclusion. International Journal
cies among different malocclusion groups. European Journal of of Orthodontics 9: 1.
Orthodontics 32: 307–312. Zerouaoui MF, Bahije L, Zaoui F and Regragui S (2014) Study of varia-
Othman S and Harradine N (2006) Tooth-size discrepancy and Bolton’s tions of the Bolton index in the Moroccan population depending on
ratios: a literature review. Journal of Orthodontics 33: 45–51. angle malocclusion class. International Orthodontics 12: 213–221.

A Systematic Review and Meta-Analysis

Uploaded by

Copyright:

Available Formats

You might also like

A Systematic Review and Meta-Analysis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Systematic Review and Meta-Analysis

Uploaded by

Copyright:

Available Formats

886322 JOO Journal of OrthodonticsMachado et al.