The minimally invasive procedure of filler injection has become a very popular and highly successful technique for achieving a youthful face.1 Because natural aging processes such as weakening of the facial muscle and tissue are contributing factors of skin aging, more people are becoming increasingly concerned about these features.1–3 Even though these procedures are safe, severe complications have been continually reported.3 , 4 Although rarely reported, patients have suffered from severe complications such as visual impairment, skin necrosis, and cerebral infarction.3 However, these complications are closely associated with vascular compromise.5
Direct intravascular filler injections with high pressure into the large artery can potentially cause retrograde displacement of filler particles and lead to occlusion of the retinal arteries, subsequently resulting in tissue necrosis, blindness, and cerebral infarction.1 , 4 , 5 To prevent vascular complications, the study and appreciation of arterial locations and their anastomoses are crucial. The arterial supply of the face primarily arises from the external and internal carotid systems, and both systems are extensively interconnected, offering numerous potential pathways for the occurrence of severe complications.6 , 7
Regarding filler injection sites, there are particular locations that have been reported to incur complications after filler injection. For example, regarding the periorbital region, which contains arterial anastomosis of both systems, injection at this danger zone should be avoided.4 According to Kim et al., injection into the supratrochlear and supraorbital arteries when performing glabellar augmentation commonly causes retrograde embolism by means of the anastomosis between the dorsal nasal and angular arteries.1 Moreover, injection of the temporal region, which contains the communication between the superficial temporal artery and the ophthalmic artery, is likely to cause occlusion of the retinal artery also.4 , 8 Although the occurrence of severe complications following injection of the upper face has been well described in the literature, the possibility of arterial injury and occurrence of severe complication following injection at the middle face has been poorly reported.
The tear trough is a common filler injection area at the middle face that serves as the dominant sign of aging at the periorbital region.9 This tear trough depression line extends from the medial canthus and runs inferolaterally, parallel to the infraorbital margin, and terminates adjacent to the midpupillary line.5 The arterial supply surrounding the tear trough deformity is associated with the inferior palpebral region; thus, the arterial supply surrounding the tear trough region possibly originates from both the infraorbital and the facial arteries. However, the facial artery provides the arterial supply to areas located superficial to the muscular and subcutaneous layers of the face.10 Moreover, Kelly et al. demonstrated that the angular artery is the primary blood supply of the paranasal region; however, this vessel terminates at the ala. The infraorbital artery also gives off its branches to supply the medial canthal region.11 Very few studies have attempted to illustrate the association between the facial arteries in relation to soft-tissue landmarks, and the association, if any, between identified arterial location and real filler injection sites remain unknown. Therefore, this study objectively evaluates the nearest arterial location relative to tear trough deformities and investigates the possibility of incurring severe complications following filler injection.
MATERIALS AND METHODS
Dissections of 30 hemifaces of 15 embalmed Thai cadavers (six male and nine female cadavers) with an average age of 77.9 years were conducted at the Chulalongkorn Medical Faculty and the King Chulalongkorn Memorial Hospital. Red latex was injected for distinct visualization of the arterial branches. Tear trough injection sites were then assigned as the anatomical landmarks at the inferior margin of the tear trough, which were located between the lateral limbus and lateral canthus (TT1), the mid-pupil level of the tear trough (TT2), and at 0.5 cm of tear trough from the medial canthus (TT3) (Fig. 1).12 Regarding the arteries in this study, they consist of the following: (1) facial artery, (2) ophthalmic artery, (3) infraorbital artery, (4) transverse facial artery, (5) palpebral branch of the infraorbital artery, (5) main branch of the facial artery, (6) palpebral branch of the maxillary artery, (7) palpebral branch of the detoured branch of the facial artery, (8) detoured branch of the facial artery, (9) palpebral branch of the main facial artery, (10) angular artery, (11) dorsal nasal artery, and (12) lateral nasal artery.
Dissection was then carefully performed on the middle face, layer by layer. Next, all arterial branches of the facial artery, ophthalmic artery, infraorbital artery, and transverse facial artery were traced and marked until reaching their respective terminal branches under a surgical microscope (M525 F20; Leica Microsystems, Singapore). Then, the arterial branches related to each landmark were observed as follows:
- The nearest artery to TT1, TT2, and TT3.
- The relationship between the location of the nearest artery and fascial tissue layers at each tear trough landmark.
The distances of the nearest artery to the landmarks were measured according to standard x and y axes. The x axis was correlated to the Frankfort horizontal line, formed by extending a horizontal line from the upper margin of the tragus along with the infraorbital rim (Fig. 2). The y axis consisted of a line perpendicular to the x axis, and was located anterior to the tragus. The nearest arteries at each tear trough landmark were first identified to measure the distance from injection sites to the arteries along the x and y axes using a digital Vernier caliper (code no. 500-196-30; Mitutoyo, Kanagawa, Japan) (Fig. 2). In addition, the external diameter of the artery was examined at all landmarks.
The location of the nearest artery was defined along the x axis by the following: −x (medial to the landmarks) and +x (lateral to the landmarks). For the y axis, locations were grouped as −y and +y, referring to inferior and superior positions to the landmarks, respectively. Moreover, x 0 and y 0 were indicated that the nearest artery was placed at the landmark. Then, the relationship between landmarks and the nearest artery was classified into eight types, as follows:
- Type I: the artery lining superomedial to landmarks (−x, +y).
- Type II: superolateral (+x, +y).
- Type III: inferomedial (−x, −y).
- Type IV: inferolateral (+x, −y).
- Type V: medial (−x, y 0).
- Type VI: lateral (+x, y 0).
- Type VII: inferior (x 0, −y).
- Type VIII: superior (x 0, +y).
Statistical analysis was performed using IBM SPSS Version 21 (IBM Corp., Armonk, N.Y.) by analyzing descriptive statistics, including mean, minimum, maximum, and standard deviations. The t test was implemented to compare measurements between sexes when the data had a normal distribution. In cases of abnormal data distribution, the Mann-Whitney U test was used to compare measurements.
The tear trough area received its blood supply from a variety of sources. In 13 cases (43.3 percent), the nearest artery to TT1 was the palpebral branch of the infraorbital artery (Fig. 3, above and center). In five cases (16.7 percent), the palpebral branch of the transverse facial artery was located closest to TT1. However, the infraorbital artery and the main branch of the facial artery were found at equal distances to TT1 in three cases (10.0 percent). The maxillary artery was found to produce the palpebral branch of the maxillary artery, and located nearest to TT1 in one case (3.3 percent), and the palpebral branch of the detoured branch of the facial artery, which was located closest to TT1 in one case. In two cases, both the detoured branch of the facial artery and the palpebral branch of the main facial artery were found closest to TT1 (Table 1 and Fig. 3, above, left).
The relationships and the distances of the nearest artery to each tear trough landmark were classified into eight types and are listed in Table 2. In most cases, at TT1, this study found that the closest artery was located inferolaterally to TT1 (type IV) in seven cases (23.3 percent); the distances were 2.79 ± 1.08 mm along the x axis and 2.88 ± 1.57 mm at the y axis. However, three types, including type III (inferomedial), type V (medial), and type VII (inferior), were found in five cases (16.7 percent). In type III, the distances were 8.83 ± 4.62 mm along the x axis and 9.96 ± 6.04 mm along the y axis; the distance was 3.47 ± 2.33 mm along the x axis in type V. The distance of type VII was 2.28 ± 1.54 mm along the y axis.
The artery located closest to TT2 was the palpebral branch of the infraorbital artery in 18 cases (60 percent) (Fig. 3, center). In five cases (16.7 percent), the main branch of the infraorbital artery was closest to TT2. In addition, the detoured branch of the facial artery was found closest to TT2 in four cases (13.3 percent). The main branch of the facial artery and the palpebral branch of the maxillary artery were discovered in two cases (6.7 percent) and one case (3.3 percent), respectively (Table 1 and Fig. 3, center, left). The relationship between TT2 and its neighboring arteries showed that most of the closest arteries were located inferior and medial to TT2 (type III) in 11 cases (36.7 percent). For the distance between the nearest artery and TT2 of type III, the distances were 4.65 ± 1.83 mm at the x axis and 7.13 ± 3.99 mm along the y axis. In type IV [six cases (20 percent)], the nearest artery was found inferolateral to TT2, with the associated distances being 4.72 ± 2.37 mm at the x axis and 3.98 ± 1.20 mm along the y axis (Table 2).
The angular artery was found to closely associate with TT3; nevertheless, the angular artery was ramified from several arterial branches: ophthalmic artery, infraorbital artery, main branch of the facial artery, and detoured branch of the facial artery. In 15 cases (50 percent), the closest artery to TT3 was the angular artery, found to branch from the ophthalmic artery (Fig. 3, below). The facial artery terminated as the angular artery and was located closest to TT3 in nine cases (30 percent), whereas the peripheral end of the detoured branch of the facial artery served as the angular artery closest to TT3 in four cases (13.3 percent). The angular artery originating from the infraorbital artery and dorsal nasal artery were located closest to TT3 in one case (3.3 percent) (Table 1 and Fig. 3, below, left). In terms of neighboring artery location, the closest arteries were located medial (type V) to TT3 in most of 25 cases (83.3 percent), with the distance measured as 4.00 ± 2.37 mm at the x axis. Type III (inferomedial) and type VI (lateral) were found in two cases (6.7 percent). The distances of type III were 1.74 ± 0.52 mm at the x axis and 0.96 ± 0.30 mm along the y axis (Table 2). In addition, the detailed relationships and the distances of the nearest artery to each tear trough landmark are shown in Figure 4.
Furthermore, this study investigated the relationships between the artery and facial tissue layers based on each tear trough landmark (Fig. 5, above); our study found that most of the nearest arteries at TT1 were located deep to the muscle, including the zygomaticus major muscle and orbital part of the orbicularis oculi muscle in 21 cases (70 percent) (Fig. 5, below). For the arterial lining of TT2, the closest artery was located at the submuscular layer of the orbicularis oculi in 16 cases (53.3 percent) (Fig. 5, below). In terms of the arterial lining plane of the artery found closest to TT3, the layer in which arteries were most commonly found was the supramuscular layer of the orbicularis oculi in 27 cases (90 percent) (Table 3 and Fig. 5, below).
The external diameter measurements of the closest arteries to each tear trough landmark are listed in Table 4. No statistical significance was found (p > 0.05) when comparing vessel diameters between sexes at all landmarks (Table 4).
Filler injection techniques used to amend tear trough deformities are performed by highly experienced clinicians.13 , 14 However, the largest concern is incurring severe complications associated with inadvertent injury to the arterial supply of the face. Therefore, the clinician should understand and appreciate the arterial anatomy of the midface in great detail. This study revealed that the arterial system surrounding the tear trough area is derived from numerous arterial branches that depend on the part of the tear trough deformity.
As a result of this study, the majority of arteries closest to TT1 and TT2 were related to the palpebral branch of the infraorbital artery in 43 percent and 60 percent of cases, respectively (Table 1). Hwang et al. demonstrated that the palpebral branch of the infraorbital artery was found in 21 of 28 orbits (75.0 percent). In addition, a single branch of the palpebral branch of the infraorbital artery was discovered in 20 of 21 orbits, and two branches of the palpebral branch of the infraorbital artery were found in one orbit. The palpebral branch of the infraorbital artery was located approximately half an eye width from the medial canthus; thus, this location was associated with tear trough deformity surrounding the mid-pupil level.15 The closest artery to TT3 was the angular artery, which ramified of the ophthalmic artery in 50 percent. Therefore, this study pointed out that the main artery of the tear trough arose from the facial artery, infraorbital artery, and ophthalmic artery. However, the presence of the arteries closest to the tear trough may relate to the patterns of the angular artery. Kim et al. found that there were four patterns of angular artery. In type I (19.3 percent), the angular artery emerges from the lateral nasal artery and projects toward the medial area of the eye. Their findings correlated with our study, with 30 percent of the closest arteries to TT3, which was the angular artery from the facial artery. For type II (31.6 percent), the facial artery produces the detouring branch and projects medially to the nasojugal and medial canthal area before terminating as the angular artery. For comparison with our findings, the closest artery to TT3 (13.3 percent), TT1 (6.7 percent), and TT2 (13.3 percent) appeared to be the detoured branch of the facial artery. In addition, the angular artery, which originates from the ophthalmic artery, was grouped in type III (22.8 percent). This occurrence was associated with the most cases in our study, in which the nearest artery to TT3 was the angular artery (50 percent). In type IV (26.3 percent), the angular artery was missing and the facial artery terminated as the lateral nasal artery.16 When this type of angular artery was found, it was assumed that the angular artery may have been branches of either the infraorbital artery or the dorsal nasal artery and was located nearest to TT3 (3.3 percent). Pilsl et al. demonstrated that the angular artery was branched by the facial artery and dorsal nasal artery in 33 hemifaces (55 percent) and nine hemifaces (15 percent), respectively. In the other 18 specimens (30 percent), no angular artery was found. In this case, the small branches of the infraorbital artery played an essential role in supplying the medial angle of the eye.17 Angular artery branches originating from the ophthalmic artery may be seriously damaged during filler injection; this anastomosis provides the filler with a retrograde channel, which could lead to distribution of the filler emboli to the central retinal artery resulting in visual loss.5 , 18
Although the arterial injury was not directly found in the angular artery, the facial artery of the external carotid system may contribute to the communication with the internal carotid system to reach the embolic pathway of filler particle.7 Kim et al. described that vision loss complications that occur when performing nasolabial fold and nasojugal groove filler injection are caused by the facial artery and angular artery and their anastomoses with neighboring arteries.16 Regarding the nearest artery of TT1 and TT2, most cases were not relative to the angular artery, but were closely associated with the palpebral branch of the infraorbital artery. In this study, we also presented that the nearest artery to TT1 was the palpebral branch of the transverse facial artery in 16.7 percent. This finding corresponds to Edizer et al., who explained that the arterial supply to the inferior marginal arcade at the lower eyelid arose from the transverse facial artery, the infraorbital artery, and the inferior medial palpebral branch of the angular artery.19 However, the arterial circulation of the tear trough may provide a lower risk of vascular complications when the anastomoses between the internal and external carotid systems is absent.
Furthermore, this study provides the concerning arterial locations regarding injection sites of the tear trough. In most cases of TT1, the arterial location has several variations, including the equal presence of type III, type V, and type VII (16.7 percent in all cases). Interestingly, the distance of type VII was found as 2.28 ± 1.54 mm at the y axis, and thus may substantially increase the chance of vascular injury if the clinician incorrectly performs injection at this site. For the location at TT2, the closest artery was located inferior and medial to the landmark (type III, 36.7 percent). The distance measurements were 4.65 ± 1.83 mm at the x axis and 7.13 ± 3.99 mm along the y axis. This area closely correlated with the palpebral branch of the infraorbital artery. Hwang et al. found that the location of palpebral branch of the infraorbital artery at the medial canthus could range between 40 and 80 percent of the eye width.15 For the location at TT3, the closest artery was the angular artery branching from the ophthalmic artery and located medial to the landmark (type V, 83.3 percent) with 4.00 ± 2.37 mm at the x axis. Likely, Marur et al. described that the location of the angular artery ranged from 6 to 8 mm medial to the medial canthus and 5 mm anterior to the lacrimal sac.7 The location of our finding pointed out that injection of the tear trough deformity lateral to the medial canthus remains safe because of the distant placement from the angular artery location. For other parameters, the diameter measurement was found to be less than 1 mm at all tear trough landmarks. Tansatit et al. recommended that the size of cannulas used should be considered based on arterial size and that, furthermore, a larger cannula should be used to avoid arterial injury, as smaller cannulas provide the greater risk of arterial laceration.20 The minute arterial branches might also be damaged by the small size of the cannula or needle during injection.
Although arterial location and diameter seem to be the main causes of arterial injury, the arterial lining planes are among the clinician’s main considerations during injection. There are different types of injection technique for various injection depths, and each technique is dependent on the clinician. Hill et al. and El-Garem explained that the tear trough should be injected with a 29-gauge needle and filler should be placed superficial to the periosteum to ameliorate the shadow of the tear trough.9 , 21 Kim et al. claimed that the injection layer is dependent on two conditions. First, the layer deep to sub–orbicularis oculi fat is selected when soft-tissue atrophy appears below the orbital retaining ligament. Second, the superficial injection into the subcutaneous layer over the orbicularis oculi is used in cases of sub–orbicularis oculi fat restoration.5 Concerning our study, most cases of the arterial lining layer at TT1 and TT2 were located in the submuscular plane of the orbicularis oculi. Filler injection into the submuscular layer may possibly increase the risk of intravascular injection. However, injection into the superficial or deep subcutaneous layers surrounding the medial canthus is likely to result in arterial injury, as our study found that the nearest artery at TT3 was located within the supramuscular plane of the orbicularis oculi.
According to the conventional filler injection technique of the tear trough, the filler is usually placed deeply on the preperiosteal plane with the bolus technique. Both the needle entry site and the filler placed target site have been preferred as the same point by inserting the needle perpendicular to the skin. Based on our findings, there was no nearest artery located in the superolateral quadrant of TT1 (Fig. 4, above, left). In addition, the nearest arteries of TT2 and TT3 were located at the superolateral quadrant in only one of 30 cases, and away from TT2 (the distance along the x axis was 3.76 mm and the distance along y axis was 7.12 mm) and TT3 (the distance along the x axis was 2.48 mm and the distance along the y axis was 11.63 mm) (Fig. 4, above, right and below). Moreover, the arteries located in the superolateral region became a peripheral branch, and had a small diameter compared with the proximal part of the arteries. To recommend the safest injection technique based on our findings, the filler is intended to be deeply injected with bolus injections using a needle over the preperiosteal plane at the TT1, TT2, and TT3 areas. These areas serve as the filler placed target point because there is no artery lining on the periosteum. Before performing the injection, the inferior orbital rim and the inferior orbital fat pad of the individual patient should be palpated and compressed to prevent the filler from being deposited into the orbit, which results in eye bags. Then, the needle entry site of this injection should be obliquely accessed at the superolateral quadrant of all landmarks at 30 to 45 degrees to the skin, and the distances of the needle insertion point should not be in excess of 5 mm of x and y axes from the TT1, TT2, and TT3 landmarks (Fig. 1). The tip of the needle should be inserted and introduced in the direction of TT1, TT2, and TT3 until it gently touches the bone, and then the filler is deposited over the periosteum (supraperiosteal plane) of the TT1, TT2, and TT3 areas. Furthermore, the mold is required to disperse the material along the tear trough and to provide the desired contour.
Considering the alternative injection, the deep injection produced unsatisfactory results by showing the deficiency of soft-tissue augmentation; the superficial filler injection of the tear trough deformity can also be added to the subcutaneous layer using a cannula. Related to our results, there was no arterial lining in the subcutaneous layer along the tear trough (Table 3). Therefore, this layer has the potential to be a safe injection plane; gentle massage is also required to reduce unexpected prominent lumps. As mentioned above, we have suggested the safe injection techniques by considering our results. Consequently, it may be of benefit by providing the most effective injection outcomes and avoiding not only arterial embolism but also ecchymosis, which have been commonly found with tear trough filler injection. Regarding the limitation of this study, these findings referred only to unique Thai embalmed cadavers as the representative Asian model; thus, comparative study between cadavers of Caucasian and Asian ethnicities should also be conducted.
This study obviously indicated that the arterial concerns of intravascular injection at the tear trough might involve not only the facial artery but also the ophthalmic artery, the infraorbital artery, and the transverse facial artery. To prevent injury of arterial branches, the injection should be performed with deep bolus injection into the supraperiosteal layer of the TT1, TT2, and TT3 areas. However, the entry site should be located superolateral to the landmarks. Moreover, these anatomical data strongly offer the arterial anatomy of the tear trough in depth and also promote the alternative tear trough injection technique to prevent arterial injury, which results in vascular severe complications.
Supported by the 100th Anniversary Chulalongkorn University Fund for Doctoral Scholarship from the Graduate School, Chulalongkorn University.
1. Kim YK, Jung C, Woo SJ, Park KH. Cerebral angiographic findings of cosmetic facial filler-related ophthalmic and retinal artery occlusion. J Korean Med Sci. 2015;30:1847–1855.
2. Fan Y, Jeong JH, You GY, Park JU, Choi TH, Kim S. An experimental model design for photoaging. J Craniofac Surg. 2015;26:e467–e471.
3. Ozturk CN, Li Y, Tung R, Parker L, Piliang MP, Zins JE. Complications following injection of soft-tissue fillers. Aesthet Surg J. 2013;33:862–877.
4. Sykes JM, Cotofana S, Trevidic P, et al. Upper face: Clinical anatomy and regional approaches with injectable fillers. Plast Reconstr Surg. 2015;136:204S–18S.
5. Kim HJ, Seo KK, Lee HK, Kim J. Clinical Anatomy of the Face for Filler and Botulinum Toxin Injection. 2016.Singapore: Springer;
6. Yang HJ, Gil YC, Lee HY. Topographical anatomy of the transverse facial artery. Clin Anat. 2010;23:168–178.
7. Marur T, Tuna Y, Demirci S. Facial anatomy. Clin Dermatol. 2014;32:14–23.
8. Lasjaunias P, Berenstein A, Doyon D. Normal functional anatomy of the facial artery. Radiology 1979;133:631–638.
9. Hill RH III, Czyz CN, Kandapalli S, et al. Evolving minimally invasive techniques for tear trough enhancement. Ophthalmic Plast Reconstr Surg. 2015;31:306–309.
10. Yang HM, Lee JG, Hu KS, et al. New anatomical insights on the course and branching patterns of the facial artery: Clinical implications of injectable treatments to the nasolabial fold and nasojugal groove. Plast Reconstr Surg. 2014;133:1077–1082.
11. Kelly CP, Yavuzer R, Keskin M, Bradford M, Govila L, Jackson IT. Functional anastomotic relationship between the supratrochlear and facial arteries: An anatomical study. Plast Reconstr Surg. 2008;121:458–465.
12. De Pasquale A, Russa G, Pulvirenti M, Di Rosa L. Hyaluronic acid filler injections for tear-trough deformity: Injection technique and high-frequency ultrasound follow-up evaluation. Aesthetic Plast Surg. 2013;37:587–591.
13. Dallara JM, Baspeyras M, Bui P, Cartier H, Charavel MH, Dumas L. Calcium hydroxylapatite for jawline rejuvenation: Consensus recommendations. J Cosmet Dermatol. 2014;13:3–14.
14. Hexsel D, Soirefmann M, Porto MD, Siega C, Schilling-Souza J, Brum C. Double-blind, randomized, controlled clinical trial to compare safety and efficacy of a metallic cannula with that of a standard needle for soft tissue augmentation of the nasolabial folds. Dermatol Surg. 2012;38:207–214.
15. Hwang K, Kim DH, Huan F, Nam YS, Han SH. The anatomy of the palpebral branch of the infraorbital artery relating to midface lift. J Craniofac Surg. 2011;22:1489–1490.
16. Kim YS, Choi DY, Gil YC, Hu KS, Tansatit T, Kim HJ. The anatomical origin and course of the angular artery regarding its clinical implications. Dermatol Surg. 2014;40:1070–1076.
17. Pilsl U, Anderhuber F, Neugebauer S. The facial artery: The main blood vessel for the anterior face? Dermatol Surg. 2016;42:203–208.
18. Tansatit T, Moon HJ, Apinuntrum P, Phetudom T. Verification of embolic channel causing blindness following filler injection. Aesthetic Plast Surg. 2015;39:154–161.
19. Edizer M, Beden U, Icten N. Morphological parameters of the periorbital arterial arcades and potential clinical significance based on anatomical identification. J Craniofac Surg. 2009;20:209–214.
20. Tansatit T, Apinuntrum P, Phetudom T. Cadaveric assessment of lip injections: Locating the serious threats. Aesthetic Plast Surg. 2017;41:430–440.
©2018American Society of Plastic Surgeons
21. El-Garem YF. Estimation of bony orbit depth for optimal selection of the injection technique to correct the tear trough and palpebromalar groove. Dermatol Surg. 2015;41:94–101.