Cosmetic

What Causes Eyelid Bags? Analysis of 114Consecutive PatientsRobert Alan Goldberg, M.D., John D. McCann, M.D., Ph.D., Danica Fiaschetti, C.O.A., andGuy J. Ben Simon, M.D.Los Angeles, Calif.

Background: The purpose of this studywas to identify the anatomical basis for per-ception of lower eyelid bags in patientsseeking aesthetic surgery and to evaluatethe cumulative contribution of differentanatomic characteristics before surgery.

Methods: The histories and photographs ofpatients whose motivation for aesthetic consul-tation was lower eyelid bags were analyzed. Sixcategoriesofanatomicbasis for the lowereyelidbags were identified. For each patient, a scorefrom 0 to 4 was given in each category. Thecumulative contribution score for each cate-gory was calculated as total points for that cat-egory for all patients, divided by the 456 totalavailable points. The authors also developed a“uniqueness score” to reflect the percentagecontribution of the worst identified anatomicproblem compared with the other problems.This was calculated for each patient as the max-imum score in one category, divided by totalpoints for that patient.

Results: A total of 114 consecutive cases wereevaluated (67 men and 47 women; mean age,52 � 11 years; age range, 23 to 76 years). Thecumulative contribution score for each ana-tomicvariablewasas follows:cheekdescentandhollow tear trough, 52 percent; prolapse of or-bital fat, 48; skin laxity and sun damage, 35;eyelid fluid, 32; orbicularis hyperactivity, 20;and triangular cheek festoon, 13. Prolapsed or-bital fat and tear trough deformity both re-ceived the higher score and were more com-mon in men as compared with women. The

average uniqueness score was 38 percent, witha range of 20 to 75 percent. No one categoryplayed a dominant role for most patients. Teartrough depression, skin laxity, and triangularmalarmoundweresignificantlymorecommonin patients older than 50 years. Linear regres-sion analysis showed that recommendationfor surgery is based on the extent of fat pro-lapse, skin elasticity, and midface descent.Significant positive correlations were foundin all six categories and in uniqueness scorescalculated by different observers (r valuesranged from 0.31 to 0.73; p � 0.001, Pearsoncorrelation), with the highest score in agree-ment with the contribution of eyelids fat (r �0.73) and skin laxity (r � 0.66); the unique-ness score correlation was r � 0.45 (p �0.001).

Conclusions: Eyelid bags do not have asingle anatomic basis. For different ana-tomic problems, different treatments arerecommended. (Plast. Reconstr. Surg. 115:1395, 2005.)

Patients frequently consult aesthetic sur-geons because of lower eyelid concerns. Com-mon complaints that we hear include eyelidbags, circles under the eye, wrinkles aroundthe eye, or a tired look.

In the past, a simplified approach to eyelidsurgery was popular. Patients who were un-happy with their lower eyelid underwent lowerblepharoplasty. Certainly, this simplified ap-proach streamlined the surgical decision-

From the Division of Orbital and Ophthalmic Plastic Surgery, Jules Stein Eye Institute, and the Department of Ophthalmology, David GeffenSchool of Medicine at UCLA. Received for publication March 16, 2004; revised June 29, 2004.

DOI: 10.1097/01.PRS.0000157016.49072.61

1395

making and decreased the requirement forlearning different types of surgery. It was effectiveonly for the patients whose problem was amena-ble to removing skin and fat, which rendered itsuboptimal much of the time. Most aesthetic sur-geons have evolved a customized approach toeyelid surgery in which the specific anatomicproblems are identified and individualized sur-gery is designed to address these problems.1–14

We studied a consecutive series of patients whopresented for consultation regarding lower eye-lid bags to characterize the anatomic featuresthat we determined to be responsible for thepatients’ aesthetic concerns.

PATIENTS AND METHODS

One hundred fourteen consecutive patientswho sought consultation for eyelid concernswere evaluated; patients were excluded if theyhad previous eyelid surgery. The authors re-viewed consultation notes and graded stan-dardized photographs taken in upgaze, down-gaze, smiling, and oblique views. We scoredpatients in six main categories of anatomiccontributions to the eyelid bags (Table I). Eachof the six anatomic problems was graded on afive-point scale with 0 � no involvement, 1 �mild, 2 � moderate, 3 � marked, and 4 �severe in each of the specific categories. Diag-nostic criteria for the six categories are sum-marized below.

Orbital Fat Prolapse

Orbital fat prolapse can be recognized by thecharacteristic shape of the orbital fat compart-ments. The central fat pad often has a cigarshape (Fig. 1). The orbital fat seems moreprominent with advancing age. It may be thatthe septum weakens, causing the orbital fat toprotrude. Alternatively, loss of volume of thecheek and subcutaneous periorbital skin maylead to unveiling of the orbital fat. The orbitalfat is defined above by its junction with the

orbicularis muscle, a contour that becomesmore apparent with loss of the subcutaneouseyelid fat. The orbital fat is defined below bythe junction of the septum at the orbital rim.The orbital fat has individual compartmentsthat can often be visualized through the skin.The lateral and central fat pockets are sepa-rated by the arcuate expansion of the inferioroblique; the central and medial pockets areseparated by the oblique muscle itself. Oftenthe separate medial, cigar shaped central, andlateral fat pockets can be individually observed,especially in upgaze.

Eyelid Fluid

Eyelid fluid has specific diagnostic features.It is worse after a salty meal or in the morning.Eyelid fluid can be limited inferiorly by theorbital rim because of the cutaneous liga-ments, but it does not show the orbital com-partmentalization of orbital fat. Eyelid fluidoften has a purplish color (Fig. 2). It does notincrease in prominence in upgaze. Eyelid fluidis a manifestation of fluid accumulation in gen-eralized fashion. The eyelid seems to have afluid sponge that accumulates fluid preferen-tially in systemic edema or local edema such asfacial allergy. It may not always be possible todistinguish the contour of a fluid bulge in thelower eyelid compared with a fat bulge. Somediagnostic features that suggest fluid include ahistory of variability; for example, increasingafter a salty meal, purplish color, and failure tofollow the contours of the demarcated fat com-plements. Orbital fat is separated by the arcu-

TABLE IAnatomic Contributions to Eyelid Bags in 114 Consecutive

Patients

CategoryCumulative

Contribution Score

Tear trough depression 238 (52%)Orbital fat prolapse 218 (48%)Loss of skin elasticity 159 (35%)Eyelid fluid 148 (32%)Orbicularis prominence 89 (20%)Triangular malar mound 61 (13%)

FIG. 1. A 55-year-old man with orbital fat grade 3 and teartrough grade 3. The central fat pad often has a cigar shape.The orbital fat has individual compartments that can often bevisualized through the skin. (Copyright 2003, Regents of theUniversity of California.)

1396 PLASTIC AND RECONSTRUCTIVE SURGERY, April 15, 2005

ate expanse of the inferior oblique laterallyand the valley of the inferior oblique medially,whereas eyelid fluid has an even contour thatdoes not respect the orbital compartments inits distribution. Compared with orbital fat, eye-lid fluid does not change much in upgaze anddowngaze.

Tear Trough Depression

The tear trough depression is an importantfeature of eyelid and midface aging. It is char-acterized by loss of subcutaneous fat with thin-ning of the skin over the orbital rim ligamentscombined with cheek descent (Fig. 3). It isoften related to the underlying bony structureand is more common in patients with eithercongenital or age-related maxillary hypoplasia.The tear trough depression may blend into thetriangular malar mound.

Loss of Skin Elasticity

Loss of skin elasticity is a critical feature ofeyelid aging, leading to rhytides, color andtexture changes, and festoon formation. Thethin skin unveils underlying irregularities in-cluding orbicularis, orbital fat, and the teartrough. Traditional blepharoplasty is not effec-tive in restoring elasticity and is not the besttreatment for skin problems.

Orbicularis Prominence

Orbicularis prominence contributes to cos-metic eyelid concerns. Although orbicularisprominence can be a feature of the youthful

eyelid, it combines with loss of skin elasticity tocontribute to dynamic and static rhytides.Many patients notice horizontal or obliquelines that are accentuated with smiling. Thesemay be more common in Asian patients.

Triangular Malar Mound

The triangular malar mound or festoon is acontour that occurs within a fluid sponge,bound by retaining ligaments along the orbitalrim and cheek (Fig. 4). Prominent triangularmalar mounds often run in families and can bevariable with an allergic component. When theskin loses elasticity, the malar mound can be-come an actual festoon. The triangular malarmound is a fluid sponge bound above by theorbital rim ligament and below by the orbitozy-gomatic ligament.

All photographs were reviewed and scoredby two masked observers (Goldberg and Si-mon), each unaware of the score given by theother observer. Correlations in grading of eachobserver in all categories and uniqueness scorewere calculated. An average score was calcu-lated between the two observers. The cumula-tive contribution score was calculated for eachof the six anatomic variables as a percentage ofall possible points. A uniqueness score was cal-

FIG. 2. Eyelid fluid is somewhat purplish, and does nothave any delineation into medial, central, and lateral pockets.It does not change much in upgaze and is present in down-gaze. If we press on the orbital rim with our finger, we can seethe fluid gather below the orbital rim. Eyelid fluid, grade 4,demonstrating the purplish color and tendency toward fes-toon formation. (Copyright 2003, Regents of the Universityof California.)

FIG. 3. Tear trough depression, grade 4. In the obliqueview we can appreciate the hollowness of the midface in thearea of the orbital rim. (Copyright 2003, Regents of theUniversity of California.)

Vol. 115, No. 5 / WHAT CAUSES EYELID BAGS? 1397

culated for each patient to reflect the percent-age contribution of the worst identified ana-tomic problem compared with the otheranatomic problems. The uniqueness score, cal-culated as the maximum score in one categorydivided by the sum of all six scores for eachindividual patient, is a measure of how impor-tant any one variable was. For example, if apatient receives a high score of 4 in one cate-gory and a low score of 1 in another category(and 0 in the others), he will have a highuniqueness score of 4/(4 � 1) � 0.8, whichimplies that one variable is the most importantcontributor for the eyelid bag in that patient.Conversely, if a patient receives a score of 2 inthree different categories (and 0 in the oth-ers), he will have a low uniqueness score of2/(2 � 2 � 2) � 0.333, suggesting that nosingle anatomic change is responsible for theeyelid bag in that patient. Photographs wereevaluated and graded by two masked observers.Correlations in grading of each observer in allcategories and uniqueness score were calcu-lated. The study complied with the policies ofthe local institute review board.

Statistical Analysis

Statistical analysis was performed using theindependent samples t test to evaluate meanscore in each category among different agegroups of patients and to evaluate the differ-ence between men and women. Pearson biva-riate correlation was used to examine the sim-ilarity of scoring between two masked observers

in each category and in uniqueness score. Lin-ear regression analysis was used to identify thecontribution of each anatomic problem on thesurgical decision. We realize that we use anarbitrary 0 to 4 scale, but we assume the changein each point in the scale is equivalent (i.e.,change from 1 to 2 is equal to change from 2 to3 or 3 to 4). If these assumptions are not met,then the probability values are approximate.Statistical analysis was carried out with Mi-crosoft Excel XP and SPSS programs.

RESULTS

One hundred fourteen consecutive caseswere evaluated (67 men and 47 women; meanage, 52 � 11 years; age range, 23 to 76 years).The cumulative contribution score for eachanatomic variable was as follows: cheek descentand hollow tear trough, 52 percent; prolapse oforbital fat, 48 percent; skin laxity and sun dam-age, 35 percent; eyelid fluid, 32 percent; orbic-ularis hyperactivity, 20 percent; and triangularcheek festoon, 13 percent (Table I); the sum ofall total points equals 456 possible points.

The average uniqueness score was 38 per-cent (�11 percent), with a range of 20 percentto 75 percent; this reflects the percentage con-tribution of the worst identified anatomicproblem compared with the other anatomicproblems (Fig. 5). There was no one categorythat played a dominant role for most patients;rather, multiple anatomic categories wereidentified as playing a role in producing theeyelid bags.

The orbital fat and tear trough were the twoanatomic problems to receive the highest cu-mulative contribution score, indicating thatthey were thought to be most important incausing the aesthetic problem. They also hadthe highest percentage of grades 3 and 4 ascompared with all other anatomic problems(31 percent and 28 percent grade 3, and eightpercent grade 4, respectively). Both anatomicalproblems were slightly more common in malesas compared with females (average score of 2.3versus 1.7 for fat prolapse and 2.4 versus 1.9 fortear trough deformity; p � 0.01 and p � 0.02,respectively).

If we compare patients under 50 years of ageand over 50 years of age, we note that teartrough depression and skin laxity were the twofactors that seemed to increase the most withincreasing age (mean score of 1.7 versus 2.4 fortear trough and 0.98 versus 1.7 for skin laxity; p� 0.001, independent samples t test); this find-

FIG. 4. This woman with chronic swelling demonstrates asignificant triangular malar mound, grade 4. (Copyright2003, Regents of the University of California.)

1398 PLASTIC AND RECONSTRUCTIVE SURGERY, April 15, 2005

ing is consistent with our understanding offacial aging (Fig. 6). Triangular malar mound(festoon) was significantly more common in

patients older than 50 years of age (0.77 versus0.36; p � 0.01). The average uniqueness scorewas similar in the two age groups.

The average uniqueness score was 0.38 (�0.12), with a minimum of 0.20 and maximumof 0.75. This suggests that for most patients,more than one feature was important. We sug-gest that the implication of this finding is thatthe surgeon must be prepared to address morethan one variable to maximally achieve thepatient’s aesthetic goals.

Linear regression analysis to evaluate thecontribution of each of the different categorieson the decision for surgery showed that recom-mendation on lower lid blepharoplasty is influ-enced by the extent of fat prolapse (� � 0.35,p � 0.001) and by the amount of skin laxity (�� 0.23, p � 0.04). Recommendation on anyother surgery, such as fat transposition, mid-face lift/implant or lower blepharoplasty, isalso influenced by the extent of tear troughdeformity.

Positive correlations were found in all sixcategories and in uniqueness scores calculatedby different observers (r values ranged from0.31 to 0.73; p � 0.001, Pearson correlation)with the highest score in agreement to thecontribution of eyelid fat (r � 0.73) and skinlaxity (r � 0.66); uniqueness score correlation

FIG. 5. Uniqueness score calculation (maximum score/sum of all six scores). (Above) This patient had moderatescores in many of the variables so that no one feature wasdominant, leading to a low uniqueness score of 0.22. (Below)This patient was scored to have particularly prominent orbitalfat and moderate tear trough with minimal contribution ofother features, so that his uniqueness score of 0.67 reflectedthe more concentrated participation of these two variables.(Copyright 2003, Regents of the University of California.)

FIG. 6. Cumulative contribution score for each category. Average value (linedbars) is shown along with values for the subpopulations under 50 years of age (solidbars) and over 50 years of age (speckled bars). (Copyright 2003, Regents of theUniversity of California.)

Vol. 115, No. 5 / WHAT CAUSES EYELID BAGS? 1399

was r � 0.45 (p � 0.001). Interobserver corre-lations greater than 0.70 are desired; this wasachieved only in the eyelid fat category.

DISCUSSION

We have found that lower eyelid bags are acomplex problem; often several anatomicchanges may contribute to patients’ percep-tion. The most common anatomic problemsthat contribute are orbital fat prolapse, lowerlid skin elasticity, and tear trough deformity;these may be more prevalent with increasingage.

We recognize that the scoring system andmethodology employed in this study are sub-jective and probably not highly reproducible;that is the nature of aesthetic surgery, which isas much art as science. We hope that the datathat we obtained will not be viewed as a defin-itive quantitative analysis of aesthetic eyelidevaluation but rather as a starting place for athoughtful approach to individualized analysisof aesthetic eyelid problems. Still, we havemanaged to show a good agreement betweentwo masked observers in different categoriesand in the uniqueness score, suggesting thatthere are consistent and identifiable anatomicfeatures in the differential diagnosis of eyelidbags. Eyelid fat was the only category with in-terobserver correlation greater than 0.70, sug-gesting that this is the anatomic change thatcan be easily identified. The other anatomicchanges may be subtler, therefore receivingdifferent scores by different observers.

In our practice we develop a customized sur-gical plan for each patient on the basis of theidentified anatomic problems. In this group of114 patients, surgeries were recommended asindicated (Table II). As expected, based on therange of identified anatomic components,many procedures, not just blepharoplasty, wererecommended. The details of all of the differ-ent surgical options for eyelid rejuvenation are

beyond the scope of this article. We will illus-trate some of the more common options todemonstrate the relationship between identifi-cation of anatomic contribution and selectionof an individualized surgical plan.

Traditional transconjunctival blepharoplastywith fat removal still plays a role for patientswith prominent orbital fat10 (Fig. 7). Fat repo-sitioning through a transconjunctival approachis an appropriate option for patients with ade-quate orbital fat and a significant tear troughdepression.11–13 Radiofrequency eyelid spongethermoplasty utilizes an insulated tungstenneedle placed transcutaneously into the fluidsponge in the lower eyelid or cheek; radiofre-quency energy is applied in closed fashion todesiccate and scar the fluid sponge14 (Fig. 8).Rejuvenation of the skin is accomplished usinga stepwise approach, including skin care pro-grams, chemical peel, and laser resurfacing(Fig. 9). Skin rejuvenation cannot compensatefor deep structural problems, however. If thereis loss of skin elasticity and cutaneous redun-dancy to the point of festoon formation, skinpinch techniques are often useful. Botulinumtoxin is useful to control orbicularis promi-nence; to reduce the risk of temporary para-lytic ectropion, conservative graded dosing isused in the lower orbicularis (Fig. 10). Whenthere is substantial deflation or descent of the

TABLE IISurgical Options Recommended Based on Anatomic

Problems

Surgical Option No.

Blepharoplasty 23Fat repositioning 26Radiofrequency eyelid sponge thermoplasty 53Laser or peel 51Botulinum toxin type A 14Midface lift with or without implant 8Fat injection 10

FIG. 7. A 45-year-old woman with grade 3 orbital fat pro-lapse before (above) and 14 months after (below) lowertransconjunctival blepharoplasty. (Copyright 2003, Regentsof the University of California.)

1400 PLASTIC AND RECONSTRUCTIVE SURGERY, April 15, 2005

malar and periorbital tissues, midface lift withor without cheek or periorbital implant is con-sidered.15–20 Fat injection is used for periorbital

volume augmentation, although the newer fill-ers such as cross-linked hyaluronic acid providea smoother contour and avoid the need fortissue harvesting (Fig. 11).

SUMMARY

The lower eyelid and midface is a focal pointof the face, and patient concerns in this areaoften lead to consultation with aesthetic sur-geons. A number of congenital and age-related

FIG. 8. A 48-year-old woman with grade 3 fluid bags before(above) and 3 months after (below) radiofrequency eyelidsponge thermoplasty. (Copyright 2003, Regents of the Uni-versity of California.)

FIG. 9. This patient with loss of skin elasticity is seen be-fore (above) and 3 months after (below) chemical peel. Somesurgeons would have recommended a skin blepharoplasty,but we believe that excising skin does not improve skin qualityand have had more success treating skin issues with chemicalpeel or resurfacing. (Copyright 2003, Regents of the Univer-sity of California.)

FIG. 10. This patient with prominent orbicularis lines(left) is improved with botulinum toxin to the lower eyelidorbicularis (right); often a small dose of 5 units spread acrossthe lower orbicularis ring is adequate to soften these orbic-ularis rolls. (Copyright 2003, Regents of the University ofCalifornia.)

FIG. 11. A 34-year-old man with grade 3 tear trough de-formity before (above) and 1 month after (below) Restylane(nonanimal stabilized hyaluronic acid, Medicis Aesthetics)injection. (Copyright 2003, Regents of the University ofCalifornia.)

Vol. 115, No. 5 / WHAT CAUSES EYELID BAGS? 1401

anatomic changes can contribute to aestheticproblems in this complex anatomic region.The better we can diagnose the contribution ofthese various anatomic components, the betterwe can design individualized surgery. Big sur-geries have big risks, and we continue to pur-sue new options for minimally invasive aes-thetic rejuvenation.

Robert A. Goldberg, M.D.Jules Stein Eye Institute100 Stein PlazaLos Angeles, Calif. 90095-7006goldberg@jsei.ucla.edu

REFERENCES

1. Rizk, S. S., and Matarasso, A. Lower eyelid blepharo-plasty: Analysis of indications and the treatment of 100patients. Plast. Reconstr. Surg. 111: 1299, 2003.

2. Flowers, R. S. Tear trough implants for correction oftear trough deformity. Clin. Plast. Surg. 20: 743, 1993.

3. Shorr, N., Hoenig, J. A., Goldberg, R. A., Perry, J. D., andShorr, J. K. Fat preservation to rejuvenate the lowereyelid. Arch. Facial Plast. Surg. 1: 38, 1999.

4. Goldberg, R. A. Lower blepharoplasty is not about re-moving skin and fat. Arch. Facial Plast. Surg. 2: 22, 2000.

5. Loeb, R. Naso-jugal groove leveling with fat tissue. Clin.Plast. Surg. 20: 393, 1993.

6. Furnas, D. W. Festoons, mounds, and bags of the eyelidsand cheek. Clin. Plast. Surg. 20: 367, 1993.

7. Freeman, M. S. Transconjunctival sub-orbicularis oculifat (SOOF) pad lift blepharoplasty: A new techniquefor the effacement of nasojugal deformity. Arch. FacialPlast. Surg. 2: 16, 2000.

8. Hoenig, J. A., Shorr, N., and Shorr, J. The suborbicu-laris oculi fat in aesthetic and reconstructive surgery.Int. Ophthalmol. Clin. 37: 179, 1997.

9. Trepsat, F. Periorbital rejuvenation combining fat graft-ing and blepharoplasties. Aesthetic Plast. Surg. 27: 243,2003.

10. Baylis, H. I., Long, J. A., and Groth, M. J. Transcon-junctival lower eyelid blepharoplasty. Ophthalmology96: 1027, 1989.

11. Goldberg, R. A., Edelstein, C., and Shorr, N. Fat repo-sitioning in lower blepharoplasty to maintain infraor-bital rim contour. Facial Plast. Surg. 15: 225, 1999.

12. Kawamoto, H. K., and Bradley, J. P. The tear “TROUF”procedure: Transconjunctival repositioning of orbitalunipedicled fat. Plast. Reconstr. Surg. 112: 1903, 2003.

13. Goldberg, R. A. Transconjunctival orbital fat reposi-tioning: Transposition of orbital fat pedicles into asubperiosteal pocket. Plast. Reconstr. Surg. 105: 743,2000.

14. Weber, P. J., Wulc, A. E., Moody, B. R., Dryden, R. M., andFoster, J. A. Electrosurgical modification of orbicu-laris oculi hypertrophy. Ophthal. Plast. Reconstr. Surg.16: 407, 2000.

15. Goldberg, R. A., Relan, A., and Hoenig, J. Relationshipof the eye to the bony orbit, with clinical correlations.Aust. N. Z. J. Ophthalmol. 27: 398, 1999.

16. Keller, G. S., Namazie, A., Blackwell, K., Rawnsley, J., andKhan, S. Elevation of the malar fat pad with a per-cutaneous technique. Arch. Facial Plast. Surg. 4: 20,2002.

17. Sasaki, G. H., and Cohen, A. T. Meloplication of themalar fat pads by percutaneous cable-suture tech-nique for midface rejuvenation: Outcome study (392cases, 6 years’ experience). Plast. Reconstr. Surg. 110:635, 2002.

18. Ramirez, O. M. Three-dimensional endoscopic mid-face enhancement: A personal quest for the idealcheek rejuvenation. Plast. Reconstr. Surg. 109: 329,2002.

19. Sullivan, S. A., and Dailey, R. A. Endoscopic subperi-osteal midface lift: Surgical technique with indicationsand outcomes. Ophthal. Plast. Reconstr. Surg. 18: 319,2002.

20. Williams, E. F., III, Vargas, H., Dahiya, R., Hove, C. R.,Rodgers, B. J., and Lam, S. M. Midfacial rejuvenationvia a minimal-incision brow-lift approach: Critical eval-uation of a 5-year experience. Arch. Facial Plast. Surg.5: 470, 2003.

1402 PLASTIC AND RECONSTRUCTIVE SURGERY, April 15, 2005