COSMETIC

The Functional Anatomy of the Deep

Facial Fat Compartments: A Detailed
Imaging-Based Investigation
Sebastian Cotofana,
Background: Injection of soft-tissue fillers into the facial fat compartments is
M.D., Ph.D.
frequently performed to ameliorate the signs of facial aging. This study was
Robert H. Gotkin, M.D.
designed to investigate the functional anatomy of the deep facial fat compart-
Konstantin Frank, M.D. ments and to provide information on the effects of injected material in rela-
Konstantin C. Koban tion to age and gender differences.
Stefan Targosinski, M.D. Methods: Forty fresh frozen cephalic specimens of 17 male and 23 female Cau-
Jonathan M. Sykes, M.D. casian body donors (mean age, 76.9 ± 13.1 years; mean body mass index, 23.6
Markus Schlager
Downloaded from http://journals.lww.com/plasreconsurg by BhDMf5ePHKbH4TTImqenVIkERnkI2IOHqT6/QnzEiMh5A6LkavRHUyONXlfPE1hm on 12/27/2018

± 5.3 kg/m2) were investigated. Computed tomographic and magnetic reso-

Alexander Schlattau, M.D. nance imaging procedures were carried out using colored contrast-enhanced
Thilo L. Schenck, M.D., materials with rheologic properties similar to commercially available soft-tissue
Ph.D. fillers. Anatomical dissections were performed to guide conclusions.
Albany and New York, N.Y.; Sacramen- Results: No statistically significant influences of age or gender were detected in
to, Calif.; Munich, Germany; Zurich, the investigated sample. Increased amounts of injected contrast agent did not
Switzerland; and Salzburg, Austria correlate with inferior displacement of the material in any of the investigated
compartments: deep pyriform, deep medial cheek, deep lateral cheek, deep
nasolabial (located within the premaxillary space), and the medial and lateral
sub–orbicularis oculi fat.
Conclusions: Increasing volume in the deep midfacial fat compartments did
not cause inferior displacement of the injected material. This underscores
the role of deep soft-tissue filler injections (i.e., in contact with the bone) in
providing support for overlying structures and resulting in anterior projec-
tion. (Plast. Reconstr. Surg. 143: 53, 2019.)

T
he number of soft-tissue filler injections to have been continuously adapted since their first
treat the signs of facial aging has increased description.3–13 Both computed tomographic3 and
substantially during the past 16 years. magnetic resonance14–19 imaging techniques have
According to a report of the American Society of been used to visualize these fat compartments, to
Plastic Surgeons, there was a 298 percent increase elucidate their role during facial aging and to pro-
in these treatments between the years 2000 and vide recommendations for safe and long-lasting
2016.1 The major goal of these minimally invasive results for soft-tissue filler procedures.
procedures is the restoration of volume within the The muscles of facial expression have been
facial fat compartments.2 shown to contribute to the formation of boundar-
Subcutaneous and deep facial fat compart- ies of the deep midfacial fat compartments [i.e.,
ments have been introduced to the scientific com- deep pyriform, deep medial cheek, deep lateral
munity, and their precise location and boundaries cheek, and deep nasolabial (located within the
premaxillary space) and the medial and lateral
From the Department of Medical Education, Albany Medical sub–orbicularis oculi fat].3–13 According to a previ-
College; private practice; Facial Plastic and Reconstructive ous magnetic resonance imaging study, these mus-
Surgery, University of California, Davis Medical Center; the cles undergo no significant age-related changes
Department for Hand, Plastic and Aesthetic Surgery, Lud-
wig Maximilian University; the Division of Thoracic Surgery,
University Hospital Zurich; and the Department of Radiol- Disclosure: None of the authors has any com-
ogy, Paracelsus Medical University Salzburg & Nuremberg. mercial associations or financial disclosures that
Received for publication January 6, 2018; accepted July 11, might pose or create a conflict of interest with the
2018. methods applied or the results presented in this ar-
Copyright © 2018 by the American Society of Plastic Surgeons ticle.
DOI: 10.1097/PRS.0000000000005080

www.PRSJournal.com 53
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • January 2019

when muscle length, thickness, or volume or fatty Healthcare, Little Chalfont, United Kingdom),
infiltration is measured15; this indicates a stable Resource ThickenUp Clear (Nestle HealthCare
location and course of the facial muscles during Nutrition GmbH, Vienna, Austria), and com-
the aging process. In a computed tomographic mercially available food coloring. The viscoelastic
imaging study,3 a constant volume of 1.0 cc was properties were compared (visually and manu-
injected into some of the deep facial fat compart- ally) to commercially available soft-tissue fillers
ments, and an inferior displacement of 1.3 to to ensure similar rheologic behavior. Different
2.6 mm was reported when comparing individuals amounts were injected into the targeted compart-
aged 54 to 75 years to those aged 75 to 104 years. ments irrespective of side, age, or sex.
Although it is widely accepted that the facial skel-
eton undergoes significant changes during aging Radiographic Imaging
(e.g., changes in facial widths and angles20–26), it is Computed tomographic scans were obtained
still unknown how these effects ultimately influ- with cadavers in the upright position, using a head
ence the overall functional anatomy of the deep rest to simulate the effects of gravity. The following
midfacial fat compartments. computed tomographic parameters were applied:
The objective of the present study was to slice thickness, 0.6 mm; field of view, 200 mm;
investigate the functional anatomy of the deep increment, 0.5 mm; voltage, 140 kV; and current,
midfacial fat compartments by using computed 400 mA/second (Figs. 1 through 4).
tomographic and magnetic resonance imag- Magnetic resonance scans were obtained with
ing methodologies combined with subsequent the cadavers in the supine position (because of
anatomical dissection. The precise anatomical spatial limitations of the head coil used), with the
location of the deep fat compartments, their following magnetic resonance imaging param-
bounding structures, the migratory potential of eters applied: sagittal T1 Vista (field of view, 270
various amounts of injected contrast material, and × 270 × 205 mm; voxel size, 0.7 × 0.7 × 0.35 mm
any age and/or gender effects on the functional with a voxel reconstruction size of 0.352 mm; sig-
anatomy of these compartments are evaluated. nal-to-noise ratio, 1.0; echo time, 18 msec; repeti-
tion time, 350 msec; 586 slices per data set); and
MATERIALS AND METHODS three-dimensional T2-weighted short inversion
time inversion recovery (field of view, 270 × 270
Sample × 204 mm; voxel size, 0.9 × 0.9 × 0.45 mm with a
We investigated 40 fresh frozen cephalic speci- voxel reconstruction size of 0.422 mm; signal-to-
mens from 17 male and 23 female Caucasian noise ratio, 1.0; echo time, 308 msec; repetition
body donors with a mean age of 76.9 ± 13.1 years time, 3000 msec; 454 slices per data set) (Fig. 5).
and a mean body mass index of 23.6 ± 5.3 kg/m2.
Specimens were screened and not included in this Anatomical Dissections
analysis if previous facial surgery or diseases dis- Cephalic specimens were dissected after the
rupted the integrity of the facial anatomy. Each imaging procedures at the Surgical Course Cen-
body donor had given informed consent while ter (Salzburg, Austria), and dissections were
alive for the use of his or her body for medical, facilitated by the previously injected colored radi-
scientific, and educational purposes. All aspects opaque material (Figs. 6 through 8).
of the study conform to the laws of the country
where the study was conducted. Analysis Strategy
On three-dimensional reconstructions, the
Injection Procedure maximal vertical and horizontal extent of each
The deep facial fat compartments investigated compartment was measured. The precise loca-
in this analysis were the deep pyriform, deep tion was evaluated in a transverse axis in relation
medial cheek, deep lateral cheek, deep nasolabial to the midline (distance between the midline and
(located within the premaxillary space), and the the medialmost and lateralmost aspects) and in a
medial and lateral sub–orbicularis oculi fat. The longitudinal axis in relation to a horizontal line at
injection procedures were transcutaneous (i.e., the level of the nasion (horizontal nasion level)
perpendicular to the skin surface) implantation (Table 1). Vertical measurements were adjusted
of colored radiopaque material using a 20-gauge, for midface height, calculated as the distance
70-mm, sharp-tip needle. The injected material between the nasion and the base of the nasal spine
consisted of Visipaque (iodixanol, 320 mg/ml; GE (as seen on sagittal computed tomographic scans),

54
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 143, Number 1 • The Deep Facial Fat Compartments

Fig. 1. Three-dimensional reconstruction of contrast-enhanced Fig. 3. Three-dimensional reconstruction of contrast-enhanced

computed tomographic scan showing the midfacial fat com- computed tomographic scan showing the deep nasolabial fat
partments: 1, deep lateral cheek; 2, medial suborbicularis oculi; (located within the premaxillary space) in blue. The angular vein
3, lateral suborbicularis oculi; 4, deep pyriform; and 5, deep (2) forms the lateral boundary, whereas the lateral nasal vein
nasolabial (located within the premaxillary space). Note that the forms the medial boundary (1). Both veins have connections to
deep medial cheek fat pad was not contrasted in this scan. the superior ophthalmic vein (3).

Fig. 2. Three-dimensional reconstruction of contrast-enhanced com-

puted tomographic scan showing the midfacial fat compartments: 1,
deep lateral cheek; 2, medial suborbicularis oculi; 3, lateral suborbicu-
laris oculi; 4, deep pyriform; and 5, deep nasolabial (located within
the premaxillary space). The location of the infraorbital foramen is Fig. 4. Three-dimensional reconstruction of contrast-enhanced
indicated by the red arrow. Note that the deep medial cheek fat was computed tomographic scan showing the deep pyriform space
not contrasted in this scan; original location is indicated by the blue (1) and the deep medial cheek fat (2). Red arrows indicate the
hash. The difference in superior extent of the deep nasolabial as com- angular artery. Note that on the right side the angular artery
pared to the deep pyriform compartment is indicated by the dotted travels within the deep pyriform space, whereas on the left side
lines, representing the difference in location above (5) and below (4) the artery courses medial to the deep medial cheek fat and is
the levator labii superiors alaeque nasi muscle. thus visible.

55
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • January 2019

the adjusted inferior boundary of each of the

investigated compartments were calculated using
bivariate correlation models (Pearson, rp). To also
adjust for influences of age and body mass index,
generalized linear models were performed using
IBM SPSS Version 23 (IBM Corp., Armonk, N.Y.).
Results were considered statistically significant at
a probability level of p ≤ 0.05 to guide conclusions.

RESULTS
In the investigated sample (n = 40), female
individuals were of lower weight (63.5 ± 16.0 kg
versus 81.6 ± 13.0 kg) and had a smaller midface
height (75.3 ± 3.0 mm versus 82.0 ± 3.9 mm)
compared to men (all p < 0.001). Increased
age correlated significantly with a lower weight
(rp = −0.461; p < 0.001) and individuals older than 76
years had a significantly reduced midfacial height
Fig. 5. Fused T1- and T2-weighted magnetic resonance imaging
(compared to those younger than 76 years): 80.87
sequences showing the deep medial cheek fat (1), the deep lat-
± 6.2 mm versus 77.51 ± 3.1 mm (p = 0.004). The
eral cheek fat (2), and the deep pyriform space (3). Red arrows
inferior margin in any of the investigated compart-
show the orbicularis oculi muscle/midcheek superficial muscu-
ments showed no significant correlation between
loaponeurotic system, providing evidence to the deep location
increasing amounts of injected “filler” and any
of the deep midfacial fat compartments.
inferior displacement of the material (all p > 0.05).
The superior margin of the deep fat compart-
and horizontal measurements were adjusted ments showed no superior displacement if there
for the transverse diameter of the cranium cal- was a muscular or ligamentous boundary (deep
culated at its maximal extent (as seen on trans- medial cheek, deep lateral cheek, deep pyriform,
verse computed tomographic scans). Correlations and medial and lateral sub–orbicularis oculi fat; all
between the amount of the injected volume and p > 0.05). An exception was the deep nasolabial

Fig. 6. Cadaveric dissection of the left face exposing the deep midfacial structures. Leva-
tor labii superioris alaeque nasi muscle, levator anguli oris muscle, angular vein, and
infraorbital neurovascular bundle emerging from the infraorbital foramen.

56
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 143, Number 1 • The Deep Facial Fat Compartments

Fig. 7. Cadaveric dissection of the midfacial fat compartments with

subsequent coloring of the deep facial fat compartments, for bet-
ter illustration. The levator labii superioris muscle has been removed
in this image to expose the deeper located fat compartments. White
ropes indicate the location of the orbicularis retaining ligament and the
zygomaticocutaneous ligament. LS, lateral sub–orbicularis oculi fat; MS,
medial sub–orbicularis oculi fat. Note that the deep medial cheek fat is
located between the angular vein (laterally) and the infraorbital fora-
men (medially).

fat, located in the premaxillary space (rp = −0.88;

p = 0.021); here, no strong muscular or strong liga-
mentous structure forms the superior margin.

Deep Nasolabial Fat Located in the Premaxillary

Space
The triangular deep nasolabial fat was iden-
tified within the premaxillary space, located
superficial to the levator labii superioris alaeque
nasi and deep to the orbital part of the orbi-
cularis oculi muscle (its superior part) and by
the midcheek superficial musculoaponeurotic
system (SMAS) (its lower part). The medial wall
is formed by the lateral nasal wall and the lat-
eral nasal vein, whereas the lateral boundary
is formed by a thin sheet of fibrous connective
tissue covering the angular vein. The loose cra-
Fig. 8. Cadaveric dissections of the retro–orbicularis oculi fat nial boundary is formed likewise by the angu-
(ROOF) and of the sub–orbicularis oculi fat (SOOF) after reflec- lar vein and its fascia, which runs inferior to
tion of the orbicularis oculi muscle (held by the four pickup for- the tear trough, whereas the inferior boundary
ceps in the image). Note that within the medial boundary of the is formed by the variable fascial fusion of the
sub–orbicularis oculi fat, the angular vein can be identified. midcheek SMAS and the levator labii superioris

57
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • January 2019

Table 1.  Absolute Vertical and Horizontal Measures between the Horizontal Nasion Level and the Midline
to the Most Superior/Inferior and Medial/Lateral Aspects of Each of the Investigated Deep Facial Fat
Compartments
Correlation between
Vertical Horizontal Distance between
Distance to Horizontal Distance to Inferior Border of
Injected Vertical Distance Superior Distance to Lateral the Compartment
Volume ± to Inferior Aspect Aspect ± SD Medial Aspect Aspect ± SD and the Nasal Root
Compartment SD (cc) ± SD (mm) (mm) ± SD (mm) (mm) and Injected Volume
Deep lateral cheek fat 0.66 ± 0.20 53.35 ± 4.98 33.99 ± 3.87 36.26 ± 3.87 53.61 ± 4.4 rp = 0.250;
p = 0.459
Deep medial cheek fat 0.66 ± 0.27 52.56 ± 4.16 34.48 ± 3.29 23.72 ± 6.12 40.67 ± 6.27 rp = 0.291;
p = 0.359
Deep nasolabial fat 0.76 ± 0.23 43.12 ± 7.02 17.4 ± 4.38 8.88 ± 1.79 23.75 ± 5.24 rp = −0.336;
p = 0.514
Deep pyriform fat 0.81 ± 0.24 51.61 ± 4.72 30.88 ± 4.44 12.65 ± 4.77 25.28 ± 5.87 rp = −0.438;
p = 0.089
Lateral SOOF 0.72 ± 0.29 35.85 ± 3.62 19.1 ± 3.97 40.45 ± 3.41 56.39 ± 3.2 rp = 0.329;
p = 0.183
Medial SOOF 0.52 ± 0.27 40.44 ± 6.18 25.17 ± 2.73 23.74 ± 5.14 42.96 ± 4.89 rp = 0.262;
p = 0.465
p, probability level; rp, correlation coefficient from Pearson correlation analyses; SOOF, sub–orbicularis oculi fat.

alaeque nasi muscle (Tables 1 and 2 and Figs. 1 angular artery was found to course within this
through 6). compartment (Figs. 1, 2, 4, 6, and 7).
The mean filling volume of this deep fat The mean injected volume was 0.81 ± 0.24
compartment was 0.76 ± 0.23, leading to a verti- cc and had a mean vertical/horizontal extent
cal extent of 25.72 ± 5.65 mm and to a horizontal of 20.73 ± 3.37/12.63 ± 2.38 mm. No significant
extent of 14.87 ± 4.38 mm. No significant correla- correlations were found between the amount of
tions were found between the volume of injected injected volume and the two-dimensional extent
material and any change in the two-dimensional of the compartment (vertical, p = 0.497; hori-
extent of this compartment (vertical, p = 0.614; zontal, p = 0.400). Increased age also showed no
horizontal, p = 0.810). Increased age also showed significant relationship to any change in either
no significant relationship to any change in posi- the position (superiormost versus inferiormost
tion (superiormost versus inferiormost boundary) boundary) or the extent (vertical versus horizon-
or extent (vertical versus horizontal) of the com- tal) of the compartment (all p > 0.05).
partment (all p > 0.05).
Deep Medial Cheek Fat Located in the Deep
Deep Pyriform Fat Located in the Deep Pyriform Medial Cheek Fat Compartment
Space The deep medial cheek fat was located
This deep fat compartment is located between between the levator anguli oris and the levator
the levator anguli oris and the levator labii supe- labii superioris alaeque nasi muscle and thus in
rioris alaeque nasi muscle. It is bounded medially the same plane as the deep pyriform fat compart-
by the lateral nasal wall (in its superior part) and ment. The medial boundary was formed by the
by the depressor septi nasi (in its inferior part). fascial sheet surrounding the infraorbital neuro-
The lateral wall is formed by the fascial sheet sur- vascular bundle emerging from the infraorbital
rounding the infraorbital neurovascular bundle foramen including the infraorbital nerve, artery,
(nerve, artery, veins) emerging from the infraor- and veins, whereas the lateral boundary was
bital foramen, separating this fat compartment formed by a thin sheet of connective tissue enclos-
from the deep medial cheek fat. The inferior ing the angular vein. The superior boundary was
boundary is formed by the fusion of the levator formed by the bony attachment of the levator labii
labii superioris alaeque nasi and the levator anguli superioris alaeque nasi muscle, and the inferior
oris muscle at the level of the nasolabial sulcus, boundary was formed by the fusion of the levator
whereas the superior boundary is formed by anguli oris and the levator labii superioris alaeque
the oblique (medial superior to lateral inferior) nasi muscle in its medial part and by the zygomati-
attachment of the levator labii superioris alaeque cus major and the transverse facial septum in its
nasi muscle to the maxilla (Tables 1 and 2). The lateral part (Table 3 and Figs. 1 through 7).

58
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 143, Number 1 • The Deep Facial Fat Compartments

Table 2.  Current Description in the Literature of the Boundaries of the Premaxillary and Deep Pyriform Space
and the Identified Boundaries Based on the Results of the Present Investigation
Premaxillary Space Deep Pyriform Space
Border Literature Proposal Literature Proposal
Superior Tear trough ligament 12
Angular vein — Bony attachment of the
levator labii superioris
alaeque nasi muscle
Inferior Pair of broad, transversely orientated Fascial fusion of the mid- — Levator anguli oris
retaining ligaments12; superior fornix of cheek SMAS and the muscle
the oral mucosa13 levator labii superioris
alaeque nasi muscle
Medial Nasal sidewall, levator labii superioris Lateral nasal wall and Depressor Lateral nasal wall and
alaeque nasi and nasalis muscle12; canine lateral nasal vein septi nasi8 depressor septi nasi
fossa of the maxilla 13
muscle
Lateral Medial pupil line, is an area of loose Angular vein Deep medial Infraorbital neurovascu-
areolar tissue approximately 5 mm12; cheek fat8 lar bundle
buccomaxillary ligaments13
Floor Levator labii superioris muscle12; canine Levator labii superioris — Levator anguli oris mus-
fossa of the maxilla13 alaeque nasi muscle cle and periosteum of
the maxilla
Roof Upper half of its roof is formed by the Orbital part of the orbicu- Deep medial Levator labii superioris
orbital part of the orbicularis oculi laris oculi muscle (in cheek fat8 alaeque nasi muscle
and the lower half is formed by the its superior part) and
midcheek SMAS12 midcheek SMAS (in its
lower part)

Table 3.  Current Description in the Literature of the Boundaries of the Deep Medial and Lateral Cheek Fat
Compartments and the Identified Boundaries Based on the Results of the Present Investigation
Deep Lateral Cheek Fat Compartment Deep Medial Cheek Fat Compartment
Border Literature Proposal Literature Proposal
Superior — Zygomaticocutaneous ORL ; oblique line of the
4,5,10
Bony attachment of the levator
ligament and/or zygo- maxilla6; zygomatic ligament7; labii superioris alaeque nasi
maticus minor muscle zygomaticus major muscle9 muscle
Inferior — Zygomaticus major SOOF10; facial vein7 Fusion of the levator anguli oris
muscle and transverse and the levator labii superio-
facial septum ris alaeque nasi muscle in its
medial part and by the zygomat-
icus major and the transverse
facial septum in its lateral part
Medial Deep medial Angular vein and deep Pyriform ligament5,10; lateral wall Infraorbital neurovascular bundle
cheek fat pad3 medial cheek fat of the nose6; facial vein7; levator
anguli oris9; pyriform and ORL4
Lateral Buccal fat pad3 Zygomaticus major Zygomaticus major and buccal Angular vein and deep lateral
muscle and transverse fat pad capsule6,10; zygomaticus cheek fat
facial septum major ; maxilla
9 5

Floor — Periosteum of the Periosteum of the maxilla5,10; Periosteum of the maxilla

maxilla Ristow space6; levator labii supe-
rioris7; levator anguli oris9
Roof Superficial Orbicularis oculi muscle Subcutaneous medial and mid- Levator labii superioris alaeque
medial cheek and the midcheek dle cheek fat compartment5,10; nasi muscle
fat pad3 SMAS orbicularis oculi muscle4 and
SMAS7; alar crease9
ORL, orbicularis retaining ligament; SOOF, sub–orbicularis oculi fat.

The mean injected volume was 0.66 ± 0.27 cc the extent (vertical versus horizontal) of the com-
and the mean vertical/horizontal extent was 18.08 partment (all p > 0.05).
± 4.07/16.95 ± 4.35 mm. No significant correla-
tions were found between the amount of injected Deep Lateral Cheek Fat Located in the Deep
volume and the two-dimensional extent of the Lateral Cheek Fat Compartment
compartment (vertical, p = 0.904; horizontal, The inverted triangular deep lateral cheek
p = 0.112). Increased age also showed no sig- fat overlies the area around the zygomatico-
nificant relationship to a change in the position maxillary suture and is in direct contact with
(superiormost versus inferiormost boundary) or the bone. The superior boundary is formed by

59
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • January 2019

the zygomaticocutaneous ligament and/or the ± 5.20/19.22 ± 4.69 mm for the medial sub–orbi-
zygomaticus minor muscle (if present), whereas cularis oculi fat and 16.74 ± 3.61/15.95 ± 3.25 mm
the medial boundary is formed by a thin layer of for the lateral sub–orbicularis oculi fat. No signifi-
connective tissue enveloping the angular vein. cant correlations were found between the amount
The lateral and inferior boundary is formed by of injected volume and the two-dimensional extent
the zygomaticus major muscle and the transverse of the compartment for the medial sub–orbicularis
facial septum. The anterior boundary is formed oculi fat (vertical, p = 0.092; horizontal, p = 0.440)
by the orbital part of the orbicularis oculi muscle or the lateral sub–orbicularis oculi fat (vertical, p =
and the midcheek SMAS. This compartment had 0.564; horizontal, p = 0.154). Once again, increased
no connections to the buccal fat pad (located in age was not related to a change in position (supe-
the masticator space) or to the fat located within riormost versus inferiormost boundary) or to an
the buccal space (Table 3 and Figs. 1, 2, 5, and 7). increase in extent (vertical versus horizontal) of the
The mean injected volume was 0.65 ± 0.20 cc compartment (all p > 0.05).
and the mean vertical/horizontal extent was 19.36
± 3.29/17.36 ± 6.15 mm. No significant correla- DISCUSSION
tions were found between the amount of injected
The results of this study summarize and expand
volume and the two-dimensional extent of the com-
the current understanding of the deep midfacial
partment (vertical, p = 0.832; horizontal, p = 0.182).
fat compartments and provide information on
Increased age showed no significant relationship to
their precise anatomical location. Furthermore,
any change in position (superiormost versus inferi-
the study describes comprehensively the boundar-
ormost boundary) or extent (vertical versus hori-
ies of the deep midfacial fat compartments based
zontal) of the compartment (all p > 0.05). on computed tomographic and magnetic reso-
nance imaging accompanied by anatomical dis-
Sub–Orbicularis Oculi Fat Located in the Sub– sections: deep pyriform, deep medial cheek, deep
orbicularis oculi fat Compartment lateral cheek, deep nasolabial (located within the
In 100 percent of the investigated cases, the premaxillary space), and the medial and lateral
sub–orbicularis oculi fat was divided into two sepa- sub–orbicularis oculi fat. The results reveal that
rate compartments: medial and lateral. The supe- the location, when estimated by the position of
rior boundary is the bilaminar orbicularis retaining the inferiormost and superiormost aspects of each
ligament, and the inferior boundary is formed of the visualized compartments, is not influenced
by the zygomaticocutaneous ligament and/or by age or by the amount of injected material. No
the zygomaticus minor muscle (if present). The significant increase in the two-dimensional extent
medial boundary extends until a vertical line pass- (vertical versus horizontal) was detected when
ing through the medial margin of the pupil and is various amounts of material were injected or with
not connected to the tear trough area. The angu- increased age.
lar vein is embedded in the medial boundary, and In our opinion, the strengths of this study are
the vein courses inferior to the tear trough toward the large sample size (40 fresh frozen cephalic spec-
the medial canthus. The lateral boundary is open imens); the upright positioning of the cephalic
and connected by means of the temporal tunnel to specimen during computed tomographic imag-
the inferior temporal compartment. This passage ing, simulating the effects of gravity; and the pre-
is superiorly framed by the lateral orbital thicken- cise visualization of each of the midfacial deep fat
ing and inferiorly by the McGregor patch—the compartments combined with anatomical dissec-
starting point of the zygomaticocutaneous liga- tions. Because of the injection of colored material
ment. The sub–orbicularis oculi fat lies on a thin during the imaging process, a direct comparison
sheet of fibrous connective tissue that is the mid- was possible in the consequent anatomical dissec-
facial extension of the superficial lamina of the tions. Doing so, the anatomical boundaries were
deep temporal fascia and is thus separated from identifiable and thus comparisons to the cur-
the prezygomatic space, which lies deep to this fat rent descriptions in the literature were possible
compartment (between fascia and periosteum) (Tables 2 through 4). However, a limitation of the
(Table 4 and Figs. 1, 2, 7, and 8). presented results is that the study was carried out
The mean injected volume was 0.52 ± 0.27 cc in cadaveric specimens, which lack systolic blood
for the medial sub–orbicularis oculi fat and 0.72 pressure and muscular activity of the facial and
± 0.29 cc for the lateral sub–orbicularis oculi fat. masticatory muscles, both of which are able to
The maximal vertical/horizontal extent was 15.27 potentially contribute to a displacement and/or

60
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 143, Number 1 • The Deep Facial Fat Compartments

Table 4.  Current Description in the Literature of the Boundaries of the Medial and Lateral Sub–Orbicularis
Oculi Fat Compartment and the Identified Boundaries Based on the Results of the Present Investigation
Medial Sub–Orbicularis Oculi Fat Compartment Lateral Sub–Orbicularis Oculi Fat Compartment
Border
Literature Proposal Literature Proposal
Superior Lateral orbital Orbicularis retaining Lateral orbital Orbicularis retaining ligament and
thickening11 ligament thickening11 lateral orbital thickening
Inferior Tear trough11 Zygomaticocutaneous Buccal fat pad3; tear Zygomaticocutaneous ligament
ligament and/or zygo- trough11 and/or zygomaticus minor muscle
maticus minor muscle
Medial Medial limbus3 Angular vein and tear Medial SOOF11 Medial SOOF
trough
Lateral Lateral SOOF11 Lateral SOOF Lateral canthus11 Open and connected by means of
the temporal tunnel to the infe-
rior temporal compartment
Floor Periosteum of the Midfacial extension of the Zygoma3 Midfacial extension of the superfi-
maxilla3 superficial lamina of the cial lamina of the deep temporal
deep temporal fascia fascia
Roof Subcutaneous nasolabial Orbicularis oculi muscle Lateral orbital compartment Orbicularis oculi muscle
and medial cheek fat3 and middle cheek fat3
SOOF, sub–orbicularis oculi fat.

breakdown of the injected soft-tissue filler. These 65 years as compared to 18- to 30-year-old indi-
influencing factors were—because of the nature viduals when measured by multiplanar computed
of this cadaveric study—not studied and thus no tomographic imaging. Pessa et al.21 reported an
conclusion was drawn on this aspect. inferior displacement of the inferior orbital rim
Our results are in line with most of the pre- of 9.3 to 10.6 mm in their study sample, rang-
vious descriptions of the deep facial fat compart- ing from 18 to 80 years of age. Comparing these
ments3–13 and provide additional clarification, results to the inferior displacement of the facial
especially when incorporating the concept of the fat compartments reported previously,3 similari-
course of the angular vein.27 Another strength ties (1.3 to 2.6 mm versus 0.5 to 2.0 mm) can be
of the present study is that we injected variable observed, and their descent can be potentially
amounts of material to test the expansibility and related more to the changes of the facial skeleton
the migratory potential of each of the investi- than to the descent of the fat compartments them-
gated midfacial fat compartments. Interestingly, selves. To account for this effect, we adjusted our
we found no significant influence of either age or measurements to facial skeleton parameters [i.e.,
the amount of volume injected on the vertical or vertical measures adjusted to the midfacial height
horizontal displacement of the visualized material (distance from the nasion to the base of the nasal
when measured by computed tomographic imag- spine), and horizontal measures adjusted to the
ing. This is contradictory to a previous computed widest transverse cranial diameter]. Our results
tomographic imaging investigation of nine fresh were able to show that in our investigated sample,
frozen cephalic specimens performed by Gierl- the midfacial height decreases with advanced age
off et al.3 that reported that an inferior displace- (range, 50 to 100 years), providing support for the
ment of the midfacial fat compartments occurred “concertina” effect previously postulated by Pessa
in the range of 1.3 to 2.6 mm when comparing at al.26 This age-related change in bone could be
individuals aged 54 to 75 years to those aged 75 responsible for the “secondary” sagging/shrink-
to 104 years. However, these results have to be ing phenomenon seen in the overlying soft tissues
interpreted with caution, as it is widely accepted (i.e., the facial fat compartments) and contribute
that changes of the facial skeleton occur during to the appearance of an aged face. Adjusting for
facial aging,20–26 and thus adjustments of facial facial skeleton parameters can potentially reveal
measurements have to be included in the analy- the true functional anatomy of the deep midfa-
ses because of these age-related alterations. The cial fat compartments; based on the results of this
results of Gierloff et al.3 were not adjusted to facial investigation, these compartments do not seem to
distances but were related to the inferior orbital become displaced inferiorly relative to the bone
rim. When comparing the values from the litera- during the aging process. Because the boundar-
ture measuring the inferior displacement of the ies of the deep nasolabial, deep pyriform, deep
inferior orbital rim during aging, Richard et al.28 medial cheek, and deep lateral cheek fat compart-
reported 0.5 to 2.0 mm in individuals aged 55 to ments are bounded by facial muscles (Tables 2

61
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Plastic and Reconstructive Surgery • January 2019

through 4), it is plausible that their location is ACKNOWLEDGMENTS

closely related to the origin and length of these The imaging part of this study received funding
muscles. A previous magnetic resonance imaging from Q-Med AB, Sweden (grant number 15092016),
study of 20 healthy Caucasian female subjects (16 and from MERZ Pharmaceuticals GmbH (grant num-
to 30 years versus 60 to 70 years)15 found no dif- ber 13072015). The authors would like to thank Katha-
ference in muscle length, thickness, volume, or rina Erlbacher for support in the imaging part of this
location of origin between the investigated age study and Hema Sundaram, M.D., for guidance during
groups; this supports the findings of our present discussions.
study—a “relative” stable location of the deep
facial fat compartments.
Our results are in line with a previous report REFERENCES
by Lambros,29 which stated that in the upper mid- 1. American Society of Plastic Surgeons. 2016 Plastic Surgery
face, a vertical descent of skin and subcutaneous Statistics. Available at: https://www.plasticsurgery.org/docu-
tissue is not necessarily a major component of ments/News/Statistics/2016/plastic-surgery-statistics-full-
report-2016.pdf. Accessed June 1, 2017.
aging; this was based on his observation of the sta- 2. Cotofana S, Fratila AA, Schenck TL, Redka-Swoboda W,
bility of orbicularis oculi wrinkles and moles as the Zilinsky I, Pavicic T. The anatomy of the aging face: A review.
same individuals aged over a time interval of 10 to Facial Plast Surg. 2016;32:253–260.
50 years. We found that, independent of the vol- 3. Gierloff M, Stöhring C, Buder T, Gassling V, Açil Y,
ume injected (range, 0.30 to 1.50 cc), no inferior Wiltfang J. Aging changes of the midfacial fat compart-
migration occurred in any of the investigated deep ments: A computed tomographic study. Plast Reconstr Surg.
2012;129:263–273.
fat compartments; this indicates a stable position 4. Sadick N, Dorizas A, Krueger N, Nassar A. The facial adipose
of the content of an injected compartment. system: Its role in facial aging and approaches to volume res-
These results support current concepts of soft- toration. Dermatol Surg. 2015;41(Suppl XXX1):S333–S339.
tissue filler injections wherein deep implantation 5. Ramanadham SR, Rohrich RJ. Newer understanding of
(i.e., in contact with the bone) provides support specific anatomic targets in the aging face as applied to
injectables: Superficial and deep facial fat compartments.
for the overlying structures, increases the projected An evolving target for site-specific facial augmentation. Plast
volume in an anterior orientation, and induces Reconstr Surg. 2015;136(Suppl):49S–55S.
a lifting effect of more inferiorly located facial 6. Wenjin W, Yun X, Ru-Lin H, et al. Facial contouring by tar-
soft tissues. This effect seems plausible as, during geted restoration of facial fat compartment volume: The
aging, a redistribution of fat from superficial fat to midface. Plast Reconstr Surg. 2017;139:563–572.
7. Cotofana S, Schenck TL, Trevidic P, et al. Midface: Clinical
visceral fat occurs,30–32 and there is a loss of total
anatomy and regional approaches with injectable fillers.
subcutaneous fat mass with consequent “deflated” Plast Reconstr Surg. 2015;136(Suppl):219S–234S.
fat compartments. Soft-tissue filler injections are 8. Surek CK, Vargo J, Lamb J. Deep pyriform space. Plast
thought to “reinflate” these volume-deficient com- Reconstr Surg. 2016;138:59–64.
partments (among other effects) and to restore 9. Surek CC, Beut J, Stephens R, Jelks G, Lamb J. Pertinent
the youthful appearance of the face. anatomy and analysis for midface volumizing procedures.
Plast Reconstr Surg. 2015;135:818e–829e.
10. Rohrich RJ, Pessa JE, Ristow B. The youthful cheek and
CONCLUSIONS the deep medial fat compartment. Plast Reconstr Surg.
The results of this cadaveric study suggest that, 2008;121:2107–2112.
11. Rohrich RJ, Arbique GM, Wong C, Brown S, Pessa JE. The
after adjustment for age-related changes of the anatomy of suborbicularis fat: Implications for periorbital
facial skeleton, no inferior displacement of the rejuvenation. Plast Reconstr Surg. 2009;124:946–951.
deep facial fat compartments occurs with increas- 12. Wong CH, Mendelson B. Facial soft-tissue spaces and retain-
ing age, when disregarding systolic blood pressure ing ligaments of the midcheek: Defining the premaxillary
and muscular activity. Increased injected volumes space. Plast Reconstr Surg. 2013;132:49–56.
13. Pilsl U, Rosmarin W, Anderhuber F. The premaxillary space:
did not cause inferior displacement of the injected A location for filler injection? Dermatol Surg. 2014;40:301–304.
“filler” material. This emphasizes the importance of 14. Gosain AK, Amarante MT, Hyde JS, Yousif NJ. A dynamic
deep, supraperiosteal implantation soft-tissue fillers analysis of changes in the nasolabial fold using magnetic
for the support of the overlying structures, anterior resonance imaging: Implications for facial rejuvenation and
projection, and a lifting effect of adjacent tissues. facial animation surgery. Plast Reconstr Surg. 1996;98:622–636.
15. Gosain AK, Klein MH, Sudhakar PV, Prost RW. A volu-

Sebastian Cotofana, M.D., Ph.D. metric analysis of soft-tissue changes in the aging midface
Albany Medical College using high-resolution MRI: Implications for facial rejuve-
47 New Scotland Avenue, MC-135 nation. Plast Reconstr Surg. 2005;115:1143–1452; discussion
Albany, N.Y. 12208 1153–1155.
[email protected] 16. Le Louarn C, Buthiau D, Buis J. Structural aging: The facial
Instagram: professorsebastiancotofana recurve concept. Aesthetic Plast Surg. 2007;31:213–218.

62
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.
 Volume 143, Number 1 • The Deep Facial Fat Compartments

17. Wysong A, Kim D, Joseph T, MacFarlane DF, Tang JY,

show, and the lateral suborbital trough deformity. Plast
Gladstone HB. Quantifying soft tissue loss in the aging Reconstr Surg. 2000;106:479–488; discussion 489–490.
male face using magnetic resonance imaging. Dermatol Surg. 25. Farkas JP, Pessa JE, Hubbard B, Rohrich RJ. The science
2014;40:786–793. and theory behind facial aging. Plast Reconstr Surg Glob Open
18. Wysong A, Joseph T, Kim D, Tang JY, Gladstone HB.
2013;1:1–8.
Quantifying soft tissue loss in facial aging: A study in 26. Pessa JE, Zadoo VP, Yuan C, et al. Concertina effect and facial
women using magnetic resonance imaging. Dermatol Surg. aging: Nonlinear aspects of youthfulness and skeletal remod-
2013;39:1895–1902. eling, and why, perhaps, infants have jowls. Plast Reconstr
19. Corey CL, Popelka GR, Barrera JE, Most SP. An analy-
Surg. 1999;103:635–644.
sis of malar fat volume in two age groups: Implications 27. Cotofana S, Steinke H, Schlattau A, et al. The anatomy
for craniofacial surgery. Craniomaxillofac Trauma Reconstr. of the facial vein: Implications for plastic, reconstructive,
2012;5:231–234. and aesthetic procedures. Plast Reconstr Surg. 2017;139:
20. Richard MJ, Morris C, Deen BF, Gray L, Woodward JA.
1346–1353.
Analysis of the anatomic changes of the aging facial skel- 28. Richard MJ, Morris C, Deen BF, Gray L, Woodward JA.

eton using computer-assisted tomography. Ophthalmic Plast Analysis of the anatomic changes of the aging facial skel-
Reconstr Surg. 2009;25:382–386. eton using computer-assisted tomography. Ophthalmic Plast
21. Pessa JE, Chen Y. Curve analysis of the aging orbital aperture. Reconstr Surg. 2009;25:382–386.
Plast Reconstr Surg. 2002;109:751–755; discussion 756–760. 29. Lambros V. Observations on periorbital and midface aging.
22. Kim SJ, Kim SJ, Park JS, Byun SW, Bae JH. Analysis of age- Plast Reconstr Surg. 2007;120:1367–1376; discussion 1377.
related changes in Asian facial skeletons using 3D vector 30. Sepe A, Tchkonia T, Thomou T, Zamboni M, Kirkland JL.
mathematics on picture archiving and communication sys- Aging and regional differences in fat cell progenitors: A
tem computed tomography. Yonsei Med J. 2015;56:1395–1400. mini-review. Gerontology 2011;57:66–75.
23. Shaw RB Jr, Kahn DM. Aging of the midface bony elements: 31. Palmer AK, Kirkland JL. Aging and adipose tissue: Potential
A three-dimensional computed tomographic study. Plast interventions for diabetes and regenerative medicine. Exp
Reconstr Surg. 2007;119:675–681; discussion 682–683. Gerontol. 2016;86:97–105.
24. Pessa JE. An algorithm of facial aging: Verification of
32. Lakowa N, Trieu N, Flehmig G, et al. Telomere length differ-
Lambros’s theory by three-dimensional stereolithography, ences between subcutaneous and visceral adipose tissue in
with reference to the pathogenesis of midfacial aging, scleral humans. Biochem Biophys Res Commun. 2015;457:426–432.

63
Copyright © 2018 American Society of Plastic Surgeons. Unauthorized reproduction of this article is prohibited.