ORIGINAL ARTICLE

Proximity of the roots of posterior teeth

to the maxillary sinus in different facial
biotypes
Maria-Carmen Costea,a Cosmina-Ioana Bondor,b Alexandrina Muntean,c M^ındra E. Badea,d
Anca-Ştefania Mesaroş,e and Anne Marie Kuijpers-Jagtmanf
Cluj-Napoca, Romania, and Nijmegen, The Netherlands

Introduction: Orthodontists consider facial growth pattern and oral function when developing a treatment plan.
Less attention is given to the relationship between the maxillary posterior teeth and the maxillary sinus. We
aimed to evaluate the relationship between the roots of the maxillary posterior teeth and the floor of the maxillary
sinus. Methods: Proximity of the roots to the maxillary sinus was scored for the left and right first and second
premolars and molars (scores, 0-3). Mean scores per patient and per tooth type were calculated. The influences
of age, sex, and facial biotype on mean scores per patient and tooth were analyzed. Results: The mean scores
per patient and the second molar scores were significantly lower in the normodivergent subjects compared with
the hypodivergent subjects, and in the hypodivergent vs the hyperdivergent groups, indicating that the hypodi-
vergent biotype had significantly fewer second molar roots into the sinus than the normodivergent and hyperdi-
vergent biotypes. Age had no effect on mean score per patient, but in the hyperdivergent group, the
second molar score increased with age, meaning that second molar roots tend to be closer to the sinus floor.
Conclusions: In a young population (7-24 years), the positions of the apices of the maxillary second molar roots
in relation to the maxillary sinus floor are associated with the facial biotype. In a hypodivergent biotype, the roots
of the second molars are located farther from the sinus floor compared with the normodivergent and hyperdiver-
gent facial patterns. (Am J Orthod Dentofacial Orthop 2018;154:346-55)

W
hen determining a treatment plan for a patient, maxilla. The maximum volume is reached in the
orthodontists consider the facial growth second decade in girls and the third decade in men,1 or
pattern and oral function. However, less atten- by the age of 20 to 25 years as described by other au-
tion is given to the relationship between the maxillary pos- thors.2,3 Evaluating the position of the roots of the
terior teeth and the maxillary sinus floor. The maxillary maxillary teeth in relation to the maxillary sinus is
sinuses are small at birth and enlarge with the growing important for a comprehensive orthodontic diagnosis
and treatment plan, especially when endodontic
treatment might be needed, severely displaced impacted
a
Private practice, Cluj-Napoca, Romania.
b
teeth are present, extractions or dental implants are
Faculty of Medicine, Department of Medical Informatics and Biostatistics, Iuliu
Haţieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania. considered, vertical control methods are required during
c
Faculty of Dentistry, Department of Pedodontics, Iuliu Haţieganu University of treatment, or orthognathic surgery is planned.4-6 Because
Medicine and Pharmacy, Cluj-Napoca, Romania.
d maxillary sinus enlargement during facial growth is
Faculty of Dentistry, Department of Preventive Dentistry, Iuliu Haţieganu Uni-
versity of Medicine and Pharmacy, Cluj-Napoca, Romania. related to the vertical increase in the alveolar process, we
e
Faculty of Dentistry, Department of Propaedeutics and Aesthetics, Iuliu Haţie- expected to find a relationship between the facial growth
ganu University of Medicine and Pharmacy, Cluj-Napoca, Romania.
f
pattern and the 3-dimensional (3D) position of the poste-
Department of Orthodontics and Craniofacial Biology, Radboud University Med-
ical Center, Nijmegen, The Netherlands rior maxillary teeth in relation to the maxillary sinus. There-
All authors have completed and submitted the ICMJE Form for Disclosure of Po- fore, the aim of this study was to relate the proximity of the
tential Conflicts of Interest, and none were reported.
roots of the maxillary posterior teeth to the floor of the
Address correspondence to: Alexandrina Muntean, Iuliu Haţieganu University of
Medicine and Pharmacy, Faculty of Dentistry, Department of Pedodontics, 33 maxillary sinus in different facial growth patterns.
Calea Moţilor Street, First floor, 400001 CJ Cluj-Napoca, Romania; e-mail,
ortoanda@yahoo.com.
Submitted, December 2016; revised, January 2017; accepted, January 2018. MATERIAL AND METHODS
0889-5406/$36.00
Ó 2018 by the American Association of Orthodontists. All rights reserved. This was a retrospective cross-sectional study based
https://doi.org/10.1016/j.ajodo.2018.01.006 on the dental charts of 1455 patients from a private
346
Costea et al 347

Fig 1. Images processed with Dolphin 3D software. Upper row: full size scans: A, 3D surface
rendering; B, lateral cephalogram reformat obtained from the full size CBCT scan; C, digitized
CBCT lateral cephalogram reformat obtained from the full size CBCT scan data. Lower row: limited field
of view CBCT scans: D, CBCT 3D surface rendering; E, lateral cephalogram reconstructed from the
limited field of view CBCT scan; F, digitized CBCT lateral cephalogram reconstructed from the limitied
field of view CBCT scan data.

dental practice in Cluj-Napoca, Romania. Partial or full-
size cone-beam computed tomography (CBCT) scans
were available for these patients. CBCT imaging was per-
formed for orthodontic and orthognathic treatment
planning and the diagnosis of temporomandibular ab-
normalies. From these charts, patients were selected
based on the following inclusion criteria: young adults,
adolescents, and children (ages, 7-25 years), in either
the mixed or permanent dentition. Exclusion criteria
were craniofacial deformities, genetic syndromes, sys-
temic diseases, previous injuries or trauma in the maxil-
lofacial region, or previous orthodontic treatment.
The ethics commission of the Iuliu Haţieganu Univer-
sity of Medicine and Pharmacy, Cluj-Napoca, Romania,
approved the study.
All CBCT images were acquired with the same i-CAT
CBCT machine (Imaging Sciences International, Hat-
field, Pa). The scanning parameters were 120 kV(p),
23.87 mA, exposure time of 10 to 20 seconds, and voxel
size of 0.4 mm. The x-ray machine was calibrated twice a
day, and the data were saved in DICOM format.
To determine the facial biotype, a lateral cephalogram
was reconstructed from the CBCT data set with no built-
in magnification. The images were processed using 3D
software (version 11.7 Premium; Dolphin Imaging,
Chatsworth, Calif). Because of the young age of the pa-
tients and since most scans had a limited field of view,
cephalometric landmarks such as sella, nasion, and other
cranial structures could not be identified on the scans
(Fig 1). However, to personalize the cephalometric anal-
ysis, the following landmarks were identified on the
lateral cephalogram: porion, orbitale, menton, gonion,
mandibular incisor tip, and mandibular incisor root
apex. The Frankfort mandibular plane angle (FMA) was
calculated as the angle formed by the intersection of
the Frankfort horizontal plane and the mandibular plane.
A normal value was considered to be 25
6 3
.7
Based on the Tweed triangle7 and the FMA angle, the
patients were divided into 3 facial biotype groups: group
A, normodivergent (FMA, 22
-28
); group B, hypodiver-
gent (FMA, \22
); and group C, hyperdivergent
(FMA, .28
).

American Journal of Orthodontics and Dentofacial Orthopedics September 2018  Vol 154  Issue 3
348 Costea et al

All images were analyzed by an orthodontist

(M.-C.C.) using the Dolphin 3D software. The observer
was blinded to each patient's name, age, sex, and the
reason for the CBCT scan. To assess the intraobserver
reliability, 30 randomly selected lateral cephalograms
were remeasured after 3 weeks by the same observer,
who was blinded to the results of the previous evalua-
tion.
The vertical relationship between each root tip of
the posterior teeth and the maxillary sinus was
analyzed on CBCT scans by another orthodontist
(A.-S.M.). The second observer was also blinded to pa-
tient information and the facial assessment of the first
observer.
For each patient, the maxillary left and right first and
second premolars, and left and right first and
second molars were analyzed. Premolars with either 1
root or 2 roots were analyzed. If a maxillary posterior
tooth was not fully erupted or not touching the
mandibular antagonist teeth, it was excluded from the
study.
Each root tip was tagged with a landmark in the
Dolphin 3D digitize/measurement menu. These land-
marks were then saved for visualization in the coronal
and axial views, and rechecked in the 3D image view
to ensure that the tip of the root was properly assessed
(Fig 2). The relationship was then categorized after a
root type classification modified from the study of
Jung and Cho8 (Fig 3), and a score from 0 to 3 was as-
signed to each root depending on its relationship with
the maxillary sinus: 0, the root of the tooth is away
from the cortical border of the sinus with a zone of
cancellous bone in between (Fig 3, A); 1, the root is
laterally projected, away from the cortical border of
the sinus (Fig 3, B); 2, the tip of the root is in contact
with the cortical border of the sinus (Fig 3, C); and 3,
the tip of the root projects into the maxillary sinus
(Fig 3, D). The most favorable score from a dental
point of view was 0, and a score of 3 was the most
unfavorable.
To determine the intraobserver reliability, 30
Fig 2. Dolphin 3D images: A, landmarks placed on every
randomly selected teeth were reassessed after 3 weeks
root apex of the maxillary posterior teeth; B, coronal slide
by the same observer blinded to the results of the previ- showing dental roots with the assessed landmarks; C, ac-
ous evaluation. curate placement of the root landmarks in the axial view.

Statistical analysis
For each patient, an average patient score (Patient In the same manner, without taking sides into ac-
score) was calculated using the following formula: count, an average tooth score (Tooth score) was calcu-
P lated for the first premolar (PM1 score), second
Patient score 5
of all root scores of the patient premolar (PM2 score), first molar (M1 score), and
total number of roots of the patient second molar (M2 score) using the following formula:

September 2018  Vol 154  Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Costea et al 349

Fig 3. Modified scores from the study of Jung and Cho8 for root tip-maxillary sinus relationship: A,
score 0: the root of the tooth is away from the cortical border of the sinus, with a zone of cancellous
bone in between; B, score 1: the root is laterally projected, away from the cortical border of the sinus;
C, score 2: the tip of the root is in contact with the cortical border of the sinus; D, score 3: the tip of the
root is projecting into the maxillary sinus.

P P
left tooth root scores 1 right tooth root scores
Tooth score 5
number of left 1 right roots

We divided our sample into age categories: children
(younger than 10 years), adolescents (10-19 years),
and young adults (over 19 years).9
All statistical analyses were performed with SPSS Sta-
tistics for Windows software (version 19.0; IBM Armonk,
NY). To determine the test-retest reliability of the assess-
ment of facial biotype and root proximity to the maxil-
lary sinus (qualitative data), the unweighted Cohen
kappa coefficient was calculated.
Data were presented as percentages or arithmetic
means and standard deviations for scores, and medians
and interquartile intervals for age, teeth per patient,
and roots per patient. The chi-square test was used to
analyze the differences between groups according to
sex. For quantitative data, we tested the normal distribu-
tion using the Kolmogorov-Smirnov test. The Kruskal-
Wallis test was used to analyze the differences between
facial biotype groups for variables that were not nor-
mally distributed: age, teeth per patient, roots per pa-
tient, Patient scores, and Tooth scores. The analysis
was followed by the post hoc test with the Bonferroni
correction.
The Mann-Whitney U test was used to analyze sex
differences in the computed average scores. Correlations
between PM1 score, PM2 score, M1 score, and M2 score
were analyzed by the Spearman coefficient of correla-
tion.
The influence of facial biotype on Tooth score and
Patient score was analyzed using a general linear regres-
sion model with fixed effects after controlling for con-
founding factors identified in univariate analysis. The
results for each factor are the adjusted coefficient, the
95% confidence interval of the same coefficient, and
the P value. In the parameter estimates, the hypodiver-
gent facial biotype group was taken as the reference
group.
Significance was set at P \0.05.
RESULTS
The final sample consisted of 128 patients (Fig 4).
The main reason for exclusion was age greater than
25 years. The unweighted Cohen kappa coefficients for
the determination of facial biotype and root assignment
between test-retest were k 5 0.91 (P \0.001) and
k 5 0.82 (P \0.001), respectively.
The distributions of age, sex, number of teeth, and
number of roots according to facial biotype are given

American Journal of Orthodontics and Dentofacial Orthopedics September 2018  Vol 154  Issue 3
350 Costea et al

Fig 4. Patient selection flow chart.

Table I. Comparisons between groups for age, sex, and numbers of teeth and roots per patient
Total Hypodivergent Normodivergent Hyperdivergent
Parameters (n 5 128) (n 5 30) (n 5 64) (n 5 34) P value
Age (y) 12 (9-17) 14 (10-19) 12 (9.5-17) 11 (9-14) 0.15*
Male, n (%) 38 (29.7) 8 (26.7) 20 (31.3) 10 (29.4) 0.90y
Teeth per patient 6.5 (2-8) 8 (4-8) 7.5 (2-8) 4 (2-8) 0.09*
Roots per patient 15 (6-18) 16 (10-18) 15 (6-18) 10 (6-18) 0.12*
Median and interquartile interval/absolute and relative frequencies are given.
*Kruskal-Wallis test; ychi-square test.

in Table I. When we compared age, sex, number of teeth,
and number of roots by facial biotype, we found no sta-
tistically significant differences between the groups.
A total of 703 teeth were examined (389 molars, 314
premolars): 186 teeth and 420 roots in the hypodiver-
gent group, 356 teeth and 819 roots in the normodiver-
gent group, and 161 teeth and 381 roots in the
hyperdivergent group. The most frequent root scores
were 0 (34.3%) and 3 (37.1%) (Table II).
Table III shows the distribution of root position
scores for all roots per tooth type for each facial pattern.
When comparing facial biotypes in post hoc analysis, we
found significant differences between the normodiver-
gent and hypodivergent biotypes for Patient score and
M2 score, and between hypodivergent and hyperdiver-
gent biotypes for M2 score (Table IV). The hypodivergent
facial biotype had significantly fewer second molar roots
into the sinus than did the normodivergent and hyperdi-
vergent biotypes.
When the correlations between Tooth scores were
analyzed, the PM2 score significantly and positively
correlated with the PM1 score (Spearman's rho [r] 5
0.29; P 5 0.01), M1 score (r 5 0.66; P \0.001), and
M2 score (r 5 0.37; P 5 0.002). M1 score and M2 score
were also correlated (r 5 0.50; P \0.001), as well as
PM1 score and M1 score (r 5 0.32; P 5 0.003).
Age was not significantly correlated with Patient
score (r 5 0.016; P 5 0.886). The Patient score was

September 2018  Vol 154  Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Costea et al
Table II. Frequencies (number and percentage) of root position scores for the relationship between the maxillary sinus
floor and maxillary posterior teeth for the facial patterns
Hypodivergent
(n 5 30)
Root score
0 180 (42.8)
1 23 (5.5)
2 108 (25.7)
3 109 (26.0)
Total n (%) 420 (100)
For explanation of the root scores, see the legend of Figure 3.
not significantly different between age categories: chil-
dren (7-9 years; n 5 32), 1.81 6 0.57; adolescents (10-
19 years; n 5 79), 1.64 6 071, and young adults (20-
24 years; n 5 16), 1.65 6 0.66 (P 5 0.623, Kruskal-
Wallis test). The PM2 score was not significantly
different between age categories (data not shown).
When we analyzed the correlation between age and
Tooth scores, only the PM2 score was significantly and-
positively correlated with age (r 5 0.27; P 5 0.041).
The Patient score did not differ significantly for male
and female subjects (1.61 6 0.68 vs 1.71 6 0.67;
P 5 0.40, Mann-Whitney test). There were also no sig-
nificant sex differences in Tooth scores.
The results of the multivariate analysis are presented
in Table V. Age was taken as a covariate in the general-
ized linear regression model analysis, with Patient score,
PM1 score, PM2 score, M1 score, and M2 score as the
dependent variables. After controlling for age, a signifi-
cant difference was found between the normodivergent
and hypodivergent facial biotype groups for Patient
score; the normodivergent group had a higher frequency
of maxillary tooth roots near the maxillary sinus than did
the hypodivergent group. A coefficient of 0.38 translates
to a 0.38-fold increase in Patient score in the tested nor-
modivergent group compared with the hypodivergent
reference group. This was also found for the M1 score
and M2 score. A significant difference was found be-
tween the hyperdivergent and hypodivergent facial
biotype groups for M2 score: the hyperdivergent group
had a higher frequency of maxillary second molar roots
near the maxillary sinus than did the hypodivergent
reference group (Table V).
Age was associated with M2 score in the multivariate
analysis (Table V); this confirms our results from the uni-
variate analysis (r 5 0.26; P 5 0.03; n 5 69).
DISCUSSION
We studied the position of the roots of the maxillary
posterior teeth in relation to the sinus floor in different
facial biotypes and found a relationship between facial
American Journal of Orthodontics and Dentofacial Orthopedics
351
Normodivergent Hyperdivergent Total number
(n 5 64) (n 5 34) of roots
n (%)
259 (31.6) 116 (30.4) 555 (34.3)
39 (4.8) 7 (1.8) 69 (4.3)
169 (20.6) 118 (31.0) 395 (24.3)
352 (43.0) 140 (36.8) 601 (37.1)
819 (100) 381 (100) 1620 (100)
biotype and the proximity of the roots of the maxillary
posterior teeth to the maxillary sinus. Earlier studies on
the relationship between maxillary posterior teeth and
the maxillary sinus floor did not consider facial biotype.
The most unfavorable root score from the dental
perspective (ie, 3) was found most frequently in the nor-
modivergent group, indicating that the roots were
generally projected into the maxillary sinus. However,
the second molar roots in the hyperdivergent group
were closer to the sinus floor than in the hypodivergent
group. A score of 3 was also predominant in the hyper-
divergent group, with scores of 0 and 1 found in fewer
than one third of patients. The second molars were
significantly closer to the sinus floor (higher M2 score)
in the hyperdivergent group than in the hypodivergent
group. For the first molars (M1 score), we found the
same results as for the second molar, but significance
was exactly on the threshold (P 5 0.05).
Furthermore, in the hyperdivergent group, the
second molar score increased with age, meaning that
the second molar roots tended to be closer to the sinus
floor. This is an important observation from an ortho-
dontic perspective; the treatment strategy for this facial
biotype implies molar intrusion, meaning further
displacement of the molars into the sinus. In contrast,
the most frequent score in the hypodivergent group
was a favorable 0, and the Patient score was lower
than in the normodivergent group, indicating that the
roots were far from the sinus floor. This allows for
more freedom in planning orthodontic treatment strate-
gies, dental implant placement, and endodontic treat-
ment procedures in patients with a hypodivergent
biotype.
The frequency of the first and second premolars near
the maxillary sinus was not associated with facial
biotype, probably because fewer premolars than molars
were scored, and a larger sample is needed to demon-
strate the difference.
The number of teeth per patient showed a decreasing
trend from the hypodivergent to the hyperdivergent pa-
tients. This trend was due to the absence of premolars,
September 2018  Vol 154  Issue 3
352 Costea et al

Table III. Frequencies (number and percentage) of root position scores per tooth/root for the relationship between
the maxillary sinus floor and maxillary posterior teeth for the facial patterns
Roots (n) Root Score Hypodivergent Normodivergent Hyperdivergent Total
First premolar (n 5 166)
1 0 13 (36.1) 18 (50) 5 (13.9) 36 (100)
2 1 (25) 2 (50) 1 (25) 4 (100)
2 Buccal 0 33 (26.4) 63 (50.4) 29 (23.2) 125 (100)
2 (0) 1 (100) (0) 1 (100)
Palatal 0 33 (27.7) 60 (50.4) 26 (21.8) 119 (100)
2 (0) 4 (57.1) 3 (42.9) 7 (100)
Second premolar (n 5 149)
1 0 16 (30.8) 25 (48.1) 11 (21.2) 52 (100)
1 (0) 2 (100) (0) 2 (100)
2 11 (24.4) 20 (44.4) 14 (31.1) 45 (100)
3 10 (31.3) 15 (46.9) 7 (21.9) 32 (100)
2 Buccal 0 1 (10) 9 (90) (0) 10 (100)
2 1 (25) 3 (75) (0) 4 (100)
3 1 (25) 3 (75) (0) 4 (100)
Palatal 0 (0) 8 (100) (0) 8 (100)
2 2 (50) 2 (50) (0) 4 (100)
3 1 (16.7) 5 (83.3) (0) 6 (100)
First molar (n 5 253)
3 Mesiobuccal 0 15 (35.7) 14 (33.3) 13 (31) 42 (100)
1 2 (40) 3 (60) (0) 5 (100)
2 23 (30.3) 31 (40.8) 22 (28.9) 76 (100)
3 19 (14.6) 78 (60) 33 (25.4) 130 (100)
Distobuccal 0 11 (36.7) 12 (40) 7 (23.3) 30 (100)
1 2 (50) 2 (50) (0) 4 (100)
2 25 (30.9) 31 (38.3) 25 (30.9) 81 (100)
3 21 (15.2) 81 (58.7) 36 (26.1) 138 (100)
Palatal 0 15 (25.9) 25 (43.1) 18 (31) 58 (100)
1 13 (37.1) 19 (54.3) 3 (8.6) 35 (100)
2 13 (19.4) 28 (41.8) 26 (38.8) 67 (100)
3 18 (19.4) 54 (58.1) 21 (22.6) 93 (100)
Second molar (n 5 134)
3 Mesiobuccal 0 10 (62.5) 6 (37.5) (0) 16 (100)
1 1 (50) 1 (50) (0) 2 (100)
2 8 (34.8) 9 (39.1) 6 (26.1) 23 (100)
3 21 (22.6) 51 (54.8) 21 (22.6) 93 (100)
Distobuccal 0 14 (56) 9 (36) 2 (8) 25 (100)
1 1 (16.7) 4 (66.7) 1 (16.7) 6 (100)
2 12 (35.3) 12 (35.3) 10 (29.4) 34 (100)
3 13 (18.8) 42 (60.9) 14 (20.3) 69 (100)
Palatal 0 19 (55.9) 10 (29.4) 5 (14.7) 34 (100)
1 4 (26.7) 8 (53.3) 3 (20) 15 (100)
2 12 (24.5) 26 (53.1) 11 (22.4) 49 (100)
3 5 (13.9) 23 (63.9) 8 (22.2) 36 (100)
Total 420 (25.9) 819 (50.6) 381 (23.5) 1620 (100)

For explanation of the root scores, see the legend of Figure 3.

but the scores of their roots were generally low, and their
absence did not influence the general score. Also, the
differences between facial patterns can be due to the dif-
ferences in the molar scores or the influence of age as a
confounding factor (controlled with statistical multivar-
iate analysis method: generalized linear model). After
controlling for age, the differences in the molar scores
remained significant (Table V).
It is difficult to compare our results with those of
earlier studies, because our study is the first in a young
population.8,10-14 Ok et al15 analyzed the relationship
between premolars and molars with the sinus floor in a
group of 10- to 20-year-olds at a dental school clinic
as part of a larger study. All other studies were performed
with older patients varying in mean age from 17.816 to
58.8 years.10

September 2018  Vol 154  Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Costea et al 353

Table IV. Mean patient scores and mean scores per tooth for the different facial patterns
Hypodivergent Normodivergent Hyperdivergent P value
Score (n 5 30) (n 5 64) (n 5 34) (Kruskal-Wallis test)
Patient score 1.41 6 0.73* 1.79 6 0.62 1.72 6 0.67 0.03
1.67 (0.74-1.88) 1.83 (1.42-2.17) 1.67 (1.43-2.11)
PM1 score 0.04 6 0.2 0.1 6 0.3 0.14 6 0.29 0.25
0 (0-0) 0 (0-0) 0 (0-0)
PM2 score 1.49 6 1.12 1.34 6 1.12 1.5 6 0.98 0.81
1.33 (1-2.5) 1.5 (0-2.17) 1.5 (1-2)
M1 score 1.79 6 0.95 2.23 6 0.75 2.05 6 0.78 0.05
1.92 (1-2.5) 2.33 (1.83-2.83) 2.33 (1.5-2.67)
M2 score 1.51 6 0.99*,y 2.24 6 0.81 2.3 6 0.58 0.01
1.5 (0.83-2.5) 2.5 (1.83-2.83) 2.33 (1.83-2.83)

Scores were compared between facial biotypes with Kruskal-Wallis test (significant P value) followed by post hoc Bonferroni correction test; arith-
metic mean 6 standard deviation, and median (25th-75th percentiles).
*P \0.05, when comparing hypodivergent with normodivergent (post hoc Bonferroni test); yP \0.05, when comparing hypodivergent with hyper-
divergent (post hoc Bonferroni test).

Table V. Influence of the facial biotype on Tooth scores and Patient scores (dependent variables) analyzed with gen-
eral linear regression model, controlled for age (in the parameter estimates, the hypodivergent facial biotype group
was the reference group)
Adjusted coefficient b

Dependent variable Independent variable b 95% CI for b P value R2

Patient score Constant 1.50 1.04-1.96 \0.001 0.056
Age (mo) 0.006 0.033-0.020 0.641
Group (normodivergent)* 0.38 0.09-0.67 0.011
Group (hyperdivergent)y 0.30 0.003-0.64 0.075
PM1 score Constant 0.172 0.102-0.447 0.216 0.030
Age (mo) 0.008 0.024-0.008 0.304
Group (normodivergent)* 0.06 0.08-0.20 0.382
Group (hyperdivergent)y 0.09 0.08-0.26 0.293
PM2 score Constant 1.26 0.06-2.47 0.040 0.008
Age (mo) 0.014 0.054-0.083 0.678
Group (normodivergent)* 0.16 0.75-0.43 0.589
Group (hyperdivergent)y 0.02 0.70-0.73 0.960
M1 score Constant 1.43 0.88-1.99 \0.001 0.064
Age (mo) 0.024 0.008-0.057 0.136
Group (normodivergent)* 0.46 0.10-0.81 0.011
Group (hyperdivergent)y 0.32 0.09-0.72 0.123
M2 score Constant 0.50 0.054-1.54 0.341 0.204
Age (mo) 0.060 0.002-0.118 0.043
Group (normodivergent)* 0.72 0.26-1.17 0.002
Group (hyperdivergent)y 0.82 0.26-1.38 0.005

*Normodivergent group compared with hypodivergent group- the reference group; yHyperdivergent group compared with hypodivergent group-
the reference group.

The size of the maxillary sinus is not constant over
time. At birth, the maxillary sinus is small and continues
to develop in lateral and inferior directions. Jun et al1
demonstrated that the volume of the maxillary sinus in-
creases until the second decade in girls and the third
decade in men, after which it decreases. Therefore, we
hypothesized that, in a young population (children, ad-
olescents, young adults), the roots of the maxillary teeth
are more closely related to the sinus floor than in older
subjects. This agrees with the findings of Ok et al,15
who showed that the frequency of scores 2 and 3 de-
creases with age, and the frequency of score 0 increases,
meaning that the roots and the sinus drift apart with age.
In this study, a score of 3 was the most frequent vertical
relationship for the molar roots, except for the palatal
root of the second molar, for which a score of 2 was

American Journal of Orthodontics and Dentofacial Orthopedics September 2018  Vol 154  Issue 3
354
more frequent. Authors of other studies11-14 have found
in older samples ranging from mean ages of 38.811 to
55.8 years13 that the most frequent relationship for
molar roots was a score of 0, indicating that the roots
of the teeth are away from the cortical border of the si-
nus with a zone of cancellous bone in between.
Facial biotype is related to the growth of the
mandible, and its rotation will affect the direction and
amount of growth at the level of the nasomaxillary com-
plex. The growth of the maxilla is a combination of pas-
sive forward and vertical displacement and active growth
in response to soft tissue stimuli.17,18 The maxilla follows
the mandibular growth pattern and responds by
periosteal vertical apposition at the level of the
dentoalveolar process.19 In hyperdivergent patients,
because of the short mandibular ramus and conse-
quently short posterior facial height, the maxillary plane
rotates, inducing a compensatory dentoalveolar mecha-
nism in an anteroposterior direction. This is supported
by the study of Sadek et al,20 who found no difference
in posterior dentoalveolar height between facial patterns
and that the dentoalveolar compensatory mechanism
acts by vertical adaptation in both arches, but only in
the frontal alveolar process. In addition, the hyperdiver-
gent facial biotype has less dense buccal cortical bone
than do the other facial biotypes.21 This may allow for
expansion of the maxillary sinus and development to-
ward the mandibular border. Thus, because of the lack
of posterior dentoalveolar compensation, molar roots
will project into the maxillary sinus. In contrast, in hypo-
divergent patients, the mandibular ramus growth will
allow appropriate vertical dentoalveolar development
and enough bone for the maxillary teeth. The muscular
forces2228 and greater osseous density21 in hypodiver-
gent patients can be mechanisms involved in the devel-
opment of the maxillary sinus and the relationship
between roots of the posterior maxillary teeth with the
sinus floor.
In this study, we found only a difference of 0.23 in
Patient scores between age categories, demonstrating
that age did not influence Patient score. This finding
suggests that the relationship between the maxillary si-
nus floor and the maxillary posterior teeth does not
change with age in a clinically significant manner, but
further studies are needed. Age-related changes in the
relationship between the posterior teeth and the maxil-
lary sinus floor in different facial biotypes should also
be studied in adults (.20-30 years, when the maxillary
sinus volume peaks).1-3 Only then will it be possible to
prove a link between facial biotype, root proximity to
the maxillary sinus floor, and maxillary sinus volume
assessed on CBCT scans and fully understand this
adaptation process.
September 2018  Vol 154  Issue 3 American Journal of Orthodontics and Dentofacial Orthopedics
Costea et al
The limitations of our study are related to the size of
the CBCT scans. To protect against radiation, limited
field-of-view scans were performed in the younger age
group, and only the FMA could be used for the assess-
ment of the facial biotype, since no other useful cepha-
lometric landmarks were available. The sample size of
our normodivergent group was approximately twice
that of both the hypodivergent and hyperdivergent
groups. Although this represents the normal distribution
in the orthodontic population, the smaller sample size of
the hypodivergent and hyperdivergent groups may have
influenced the results. The number of roots per patient
in the hyperdivergent group was smaller than in the
other 2 groups; thus, our results regarding this group
had less power.
If the difference between 2 molars with a 3-3-3
(average, 3) score and a 1-1-1 (average,1) score is
obvious, when we have differential molar root scores
such as 0-0-3 (average 5 1) and 1-1-1 (average 5 1),
the difference is not obvious. Which is better? The ques-
tion remains unresolved, and it depends on the planned
orthodontic movement of the tooth (vertical or lateral).
The 3D position of the roots of the maxillary posterior
teeth needs to be considered from the beginning of
the orthodontic treatment.29 For example, if a score 1
type of root is to be intruded, the chances of penetrating
the buccal cortical bone and pushing the roots out of its
bony margins are high. Maxillary posterior tooth intru-
sion30,31 and space closure through the maxillary sinus
are difficult and limited.32,33 The degree of achievable
intrusion through the maxillary sinus can sometimes
be the difference between conventional orthodontic
treatment and orthognathic surgery.30,34 If a score 2 or
3 type of root is to be moved through the maxillary
sinus, the design of the orthodontic appliance and the
applied forces need to be adapted to the clinical
situation.3133
CONCLUSIONS
In a young population (7-24 years), the position of
the apices of the maxillary second molar roots in rela-
tion to the maxillary sinus floor is associated with the
facial biotype. In a hypodivergent biotype, the roots
of the molars (especially the second molar) are
located farther from the sinus floor compared with
the normodivergent and hyperdivergent facial pat-
terns. A close relationship often exists between the
maxillary posterior teeth and sinus floor. Careful
analysis of the root-sinus floor relationship of the
maxillary posterior teeth before orthodontic treat-
ment is recommended because it may change the
treatment plan.
Costea et al
REFERENCES
1. Jun BC, Song SW, Park CS, Lee DH, Cho KJ, Cho JH. The analysis
of the maxillary sinus aeration according to aging process; vol-
ume assessment by 3-dimensional reconstruction by high-
resolution CT scanning. Otolaryngol Head Neck Surg 2005;132:
429-34.
2. Baweja S. Study of age related changes of maxillary air sinus from
its anteroposterior, transverse and vertical dimensions using
computerized tomographic (CT) scan. Int J Biomed Res 2013;4:
21-5.
3. Ariji Y, Kuroki T, Moriguchi S, Ariji E, Kanda S. Age changes in the
volume of the human maxillary sinus: a study using computed to-
mography. Dentomaxillofac Radiol 1994;23:163-8.
4. Larson B. Cone-beam computed tomography is the imaging tech-
nique of choice for comprehensive orthodontic assessment. Am J
Orthod Dentofacial Orthop 2012;14:402-11.
5. Aps JK. Cone beam computed tomography in paediatric dentistry:
overview of recent literature. Eur Arch Paediatr Dent 2013;14:
13-40.
6. Georgescu CE, Rusu MC, Sandulescu M, Enache AM, Didilescu AC.
Quantitative and qualitative bone analysis in the maxillary lateral
region. Surg Radiol Anat 2012;34:551-8.
7. Tweed CH. The Frankfort-mandibular incisor angle (FMIA) in or-
thodontic diagnosis, treatment planning and prognosis. Angle Or-
thod 1954;24:121-69.
8. Jung YH, Cho BH. Assessment of the relationship between the
maxillary molars and adjacent structures using cone beam
computed tomography. Imaging Sci Dent 2012;42:219-24.
9. United Nations Children's Fund. The state of the world's children
2011. Adolescence—an age of opportunity. New York: UNICEF;
2011.
10. von Arx T, Fodich I, Bornstein MM. Proximity of premolar roots to
maxillary sinus: a radiographic survey using cone-beam computed
tomography. J Endod 2014;40:1541-8.
11. Kilic C, Kamburoglu K, Yuksel SP, Ozen T. An assessment of the
relationship between the maxillary sinus floor and the maxillary
posterior teeth root tips using dental cone-beam computerized to-
mography. Eur J Dent 2010;4:462-7.
12. Sharan A, Madjar D. Correlation between maxillary sinus floor
topography and related root position of posterior teeth using
panoramic and cross-sectional computed tomography imaging.
Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2006;102:
375-81.
13. Kwak HH, Park HD, Yoon HR, Kang MK, Koh KS, Kim HJ. Topo-
graphic anatomy of the inferior wall of the maxillary sinus in Ko-
reans. Int J Oral Maxillofac Surg 2004;33:382-8.
14. Eberhardt JA, Torabinejad M, Christiansen B. A computed tomo-
graphic study of the distances between the maxillary sinus floor
and the apices of the maxillary posterior teeth. Oral Surg Oral
Med Oral Pathol 1992;73:345-6.
15. Ok E, Gungor E, Colak M, Altunsoy M, Nur BG, Aglarci OS. Evalu-
ation of the relationship between the maxillary posterior teeth and
American Journal of Orthodontics and Dentofacial Orthopedics
355
the sinus floor using cone-beam computed tomography. Surg Ra-
diol Anat 2014;36:907-14.
16. Ahn NL, Park HS. Differences in distances between maxillary pos-
terior root apices and the sinus floor according to skeletal pattern.
Am J Orthod Dentofacial Orthop 2017;152:811-9.
17. Profitt WR, Fields HW. Occlusal forces in normal and long-faced
children. J Dent Res 1983;62:571-4.
18. Enlow DH, Hans MG. Essentials of facial growth. Philadelphia: W.
B. Saunders; 1996. p. 211-9.
19. Bj€ork A. Sutural growth of the upper face studied by the implant
method. Acta Odontol Scand 1966;24:109-27.
20. Sadek MM, Sabet NE, Hassan IT. Alveolar bone mapping in sub-
jects with different facial dimensions. Eur J Orthod 2015;37:
194-201.
21. Ozdemir F, Tozlu M, Germec-Cakan D. Cortical bone thickness of
the alveolar process measured with cone-beam computed tomog-
raphy in patients with different facial types. Am J Orthod Dento-
facial Orthop 2013;143:190-6.
22. Ingervall B, Minder C. Correlation between maximum bite force
and facial morphology in children. Angle Orthod 1997;67:415-24.
23. Ingervall B, Thilander B. Relation between facial morphology and
activity of the masticatory muscles. J Oral Rehabil 1974;1:131-47.
24. Braun S. A study of maximum bite force during growth and devel-
opment. Angle Orthod 1996;66:261-4.
25. Bakke M. Bite force and occlusion. Semin Orthod 2006;12:120-6.
26. Bakke M, Tuxen A, Vilmann P, Jensen BR, Vilmann A, Toft M. Ul-
trasound image of human masseter muscle related to bite force,
electromyography, facial morphology and occlusal factors. Scand
J Dent Res 1992;100:164-71.
27. Raadsheer MC, Kiliaridis S, Van Eijden TM, Van Ginkel FC, Prahl-
Andersen B. Masseter muscle thinkness in growing individuals and
its relation to facial morphology. Arch Oral Biol 1996;41:323-32.
28. Kiliaridis S. The importance of masticatory muscle function in den-
tofacial growth. Semin Orthod 2006;12:110-9.
29. Wehrbein H, Bauer W, Wessing G, Diedrich P. The effect of the
maxillary sinus floor on orthodontic tooth movement. Fortschr
Kieferorthop 1990;51:345-51.
30. Yao CC, Lee JJ, Chen HY, Chang ZC, Chang HF, Chen YJ. Maxillary
molar intrusion with fixed appliances and mini-implant anchorage
studied in three dimensions. Angle Orthod 2005;75:754-60.
31. Re S, Cardaropoli D, Corrente G, Abundo R. Bodily tooth move-
ment through the maxillary sinus with implant anchorage for sin-
gle tooth replacement. Clin Orthod Res 2001;4:177-81.
32. Melsen B. Limitations in adult orthodontics. In: Melsen B, editor.
Current controversies in orthodontics. Chicago: Quintessence;
1991. p. 147-80.
33. Park JH, Tai K, Kanao A, Takagi M. Space closure in the maxillary
posterior area through the maxillary sinus. Am J Orthod Dentofa-
cial Orthop 2014;145:95-102.
34. Kuroda S, Katayama A, Takano-Yamamoto T. Severe anterior
open-bite case treated using titanium screw anchorage. Angle Or-
thod 2004;74:558-67.
September 2018  Vol 154  Issue 3