A Validated Grading Scale for Crow’s Feet

ALASTAIR CARRUTHERS, MD, JEAN CARRUTHERS, MD,y BHUSHAN HARDAS, MD,z MANDEEP KAUR MD,
MS,z ROMAN GOERTELMEYER, PHD,z DEREK JONES, MD,y BERTHOLD RZANY, MD, PHD,z JOEL COHEN,
MD,J MARTINA KERSCHER, MD, PHD,# TIMOTHY CORCORAN FLYNN, MD, COREY MAAS, MD,yy
GERHARD SATTLER, MD,zz ALEXANDER GEBAUER, MD,z RAINER POOTH, MD,z KATHLEEN MCCLURE,
RN,z ULLI SIMONE-KORBEL,z AND LARRY BUCHNERyy

OBJECTIVE To develop the Crow’s Feet Grading Scale for objective quantification of the severity of
lateral canthal lines and to establish the reliability of this photonumeric scale for clinical research and
practice.
MATERIALS AND METHODS A 5-point photonumeric rating scale was developed to objectively quantify
the severity of lateral canthal lines at rest and at maximum contracture of the orbicularis oculi. Nine
experts rated photographs of 35 subjects, twice, with regard to the aspect crow’s feet in comparison
with morphed images. Inter- and intrarater variability was assessed by computing intraclass correlation
coefficients.
RESULTS The agreement between the experts was significantly high. Furthermore, the test–retest
correlation coefficients were high for each expert after an overnight interval, demonstrating low
inter- and intraevaluator variability.
CONCLUSION The 5-point photonumeric scale generated spans the severity of the type of crow’s feet
for which patients most commonly seek correction. The scale is well stratified for consistent rating.
Drs. Alastair Carruthers, Jean Carruthers, Derek Jones, Berthold Rzany, Joel Cohen, Martina Kerscher,
Timothy Corcoran Flynn, Corey Maas, and Gerhard Sattler are paid consultants to Merz Pharmaceuticals.
Dr. Bhushan Hardas, Dr. Mandeep Kaur, Dr. Roman Goertelmeyer, Dr. Alexander Gebauer, Dr. Rainer Pooth,
Kathleen McClure, and Ulli Simone-Korbel are employees of Merz Pharmaceuticals. Larry Buchner is an
employee of Canfield Scientific.

T he periorbital area is one of the first places

on the face to show signs of aging, in the
form of lateral canthal lines or crow’s feet.1 As
with other facial wrinkles, the gradual loss and
disorganization of collagen and elastin fibers,
connective tissue that provides underlying support
for skin, largely cause these rhytides.2,3 Genetic
differences may alter the rate of this normal aging
process.3 In addition, environmental toxins,
predominantly ultraviolet light exposure2 and
cigarette smoking,4 can accelerate aging of the
skin. With the decrease in skin elasticity, repeated
facial expressions can become permanent lines.3
Contraction of the orbicularis oculi, the muscles
that control facial expressions and closure of eyelids,
can accentuate the appearance of crow’s feet.5

Department of Dermatology and Skin Sciences, University of British Columbia, Vancouver, British Columbia, Canada;
y
Department of Ophthalmology and Visual Sciences, University of British Columbia, Vancouver, British Columbia,
Canada; zResearch and Development, Merz Pharmaceuticals, Greensboro, North Carolina; yDepartment of
Dermatology, University of California at Los Angeles, California; zCharitéFUniversitätsmedizin Berlin, Klinik für
Dermatologie, Division of Evidence Based Medicine (dEBM), Berlin, Germany; JAboutSkin Dermatology and
DermSurgery, Englewood, and Department of Dermatology, University of Colorado, Denver, Colorado; #Division of
Cosmetic Sciences University of Hamburg, Germany; Department of Dermatology, University of North Carolina at
Chapel Hill, and Cary Skin Center, Cary, North Carolina; yyDivision of Facial Plastic and Reconstructive Surgery,
University of California at San Francisco, California; zzRosenparkklinik GmbH, Darmstadt, Germany; yyCanfield
Scientific, Fairfield, NJ

& 2008 by the American Society for Dermatologic Surgery, Inc.
Published by Wiley Periodicals, Inc.

ISSN: 1076-0512
Dermatol Surg 2008;34:S173–S178
DOI: 10.1111/j.1524-4725.2008.34367.x

S173
 VA L I D AT E D G R A D I N G S C A L E F O R C R O W ’ S F E E T
In addition to botulinum toxin-A injections,
crow’s feet may be treated with temporary or
semipermanent injectable dermal fillers such as
allogeneic or xenogeneic collagen, hyaluronic acid,
and autologous fat transfer cells.6 Fillers can
replace volume lost during the normal aging process,
which results in a more youthful facial appearance.7
Skin resurfacing with lasers, such as the
erbium-doped yttrium aluminium garnet laser,8,9
and topical agents, such as retinoic acids, glycolic
acids,10 and lactic acids, are other common
methods to minimize crow’s feet and other facial
wrinkles.
There is no validated standard for objectively
quantifying the severity of facial wrinkles in the
clinical setting. There are subjective, quantitative
rating scales that measure facial skin changes in
general11 and crow’s feet specifically.12 Numerous
objective, quantitative rating scales that measure
three-dimensional details of the skin surface use skin
replicas13,14 and optical15,16 or computer-assisted
analyses.17 Scale information validated by Hund and
colleagues18 and Carruthers and Carruthers19 was
also used for the rationale. The significant
equipment requirements for these objective scales
limit their utility in clinical practice.
A MEDLINE search using the non-Medical Subject
Heading terms lateral canthal lines, rhytides or
wrinkles; periorbital or eyes; and validated scale
revealed one study of relevance. The study examined
two scales used to determine the clinical severity of
crow’s feet: one with the face at rest and one with the
face at maximum smile.18 The study established
interobserver and intraobserver reproducibility of
these 4-point scales,18 but the authors did not
examine scale validity and responsiveness.
We believe there is a need for an objective, quanti-
tative rating scale for evaluating the severity of
crow’s feet and monitoring the response to treat-
ment. Such a scale could establish guidelines in the
clinical setting and provide a standard, objective
evaluation of clinical trial outcomes. In this article,
S174 D E R M AT O L O G I C S U R G E RY
we present the Crow’s Feet Grading Scale and the
clinimetric evaluation that establishes the reliability
of the photonumeric scale. This scale provides the
tool for objective quantification of the severity
of crow’s feet at rest (static) and with expression
(dynamic).
Methods
Assessment Scale
The Crow’s Feet Grading Scale (Figure 1A) is a 5-
point photonumeric rating scale developed to
objectively quantify the severity of crow’s feet. This
scale was applied to two separate evaluations of
crow’s feet: at rest (static) and with expression
(dynamic). The scale ratings are 0 for no wrinkles,
1 for very fine wrinkles, 2 for fine wrinkles, 3 for
moderate wrinkles, and 4 for severe wrinkles.
A clinician was involved in the development of
morphing methodology and the description of
the scales. Approximately 50 images (per scale
validation set) were selected from the photographic
database of 100 subjects based on quality and
equal distribution across each representative scale.
Using a standardized php-based computer
randomization program, 35 images per target area
or set were randomly selected from the 50 for
final inclusion in the pool. The pictures were
taken with reference to a standard smile with
identical cheek fullness as the evidence for the
smile consistency.
Validation of Assessment Scale by Experts
Nine experts in the field of aesthetic medicine each
rated 35 cases presented as photographs of real
persons (Figure 1B) with regard to the aspect ‘‘crow’s
feet’’ in comparison with morphed images. The
morphed images showed a step-wise variation of the
defined aspect. After an initial training session and
discussion of the practice cases, each expert inde-
pendently performed the assessment of the 35 cases
on each of 2 consecutive days. The data were double
entered into a database and checked for correctness.
 CARRUTHERS ET AL

Figure 1. Crow’s Feet Grading Scale. Rated photonumeric scale scores for resting (static, first row) and hyperkinetic
(dynamic, second row) crow’s feet, (A) morphed images, (B) untouched validation images.

34:S2:NOVEMBER 2008 S175

 VA L I D AT E D G R A D I N G S C A L E F O R C R O W ’ S F E E T

TABLE 1. Descriptive Statistics for Static Crow’s for the ratings per day (Tables 1 and 2). The stability
Feet of ratings was visualized as bivariate scatter plots
Day 1 2 (bubble plots). A bubble plot is a way of representing
the relationship between variables on a scatter plot.
Mean 1.88 1.99
Median 2 2
Observations on two variables are plotted in the
STD 1.23 1.24 usual way, using circles as symbols; the radii of the
Min 0 0 circles are made proportional to the associated val-
Max 4 4 ues for the third variable. The frequency of the given
scores determine the size of the bubbles, which are
TABLE 2. Descriptive Statistics for Dynamic ideally located along the diagonal.
Crow’s Feet

Day 1 2 Validity and Reliability

Mean 1.96 2.03
Median 2 2 For each aspect and time point, intraclass correlation
STD 1.25 1.31 coefficients (ICCs) were calculated according to the
Min 0 0
Max 4 4
model of random selection of cases and raters (0–0.3
is poor, 0.4–0.6 is moderate,  0.7 is high).20,21 The
appropriate ICC under these assumptions will be the
Statistical Analyses Shrout-Fleiss random set reliability coefficient,
which is used when the same raters rate all cases and
Descriptive statistics (standard deviation, mean, the raters are regarded to be a random set of all
median, maximum, and minimum) were calculated possible raters. The Pearson correlation coefficient

Figure 2. Bubble plots of all experts for static crow’s feet.

S176 D E R M AT O L O G I C S U R G E RY

Figure 3. Bubble plots of all experts for dynamic crow’s feet.
was calculated as a test–retest reliability estimator
for stability of rating per expert after an overnight
interval over all cases for each rater.
Results
The results from the training session are not included
in the evaluation.
The ICCs (static crow’s feet: first rating 0.893,
second rating 0.882; dynamic crow’s feet: first rating
0.879, second rating 0.892) were calculated based
on 35 cases rated by nine experts. As can be seen
from the ICCs, the agreement between the experts is
considerable. Furthermore the test–retest correlation
coefficients (static crow’s feet: minimal 0.904, max-
imal 0.968; dynamic crow’s feet: minimal 0.888,
maximal 0.951) were high for each expert after an
overnight interval.
CARRUTHERS ET AL
Bubble plots for each expert and view are given
in Figures 2 and 3 for comparing the bivariate
frequencies of scores between the experts. The fre-
quency of the given scores determines the size of the
bubbles. As can be seen from Figures 2 and 3, a good
intraindividual correlation was observed in both
views.
Discussion
The results from this exercise in aesthetic medicine
for comparative judgment of real-life photographs
against morphed images associated with a 5-point
scale resulted in high ICCs. The statistical model
used for these estimates assumes a random subset of
raters from a population of experts in the aesthetic
medicine field. The coefficients can therefore be
regarded as conservative estimates of interrater
reliability. Test–retest reliability coefficients were
high, indicating sufficiently high stability of the
34:S2:NOVEMBER 2008 S177
 VA L I D AT E D G R A D I N G S C A L E F O R C R O W ’ S F E E T
ratings after an overnight interval. The expert
raters used the morphed images to judge photo-
graphs of live patients. Morphed images, although
standardized to a site-specific area, do not translate
clinically to the multiple physical changes that occur
in an aging face. The photographic evaluation and
live patient evaluation are distinct, and thus
should a clinical trial use this validated scale, the
evaluations in the clinical trial should be performed
on standardized photographs, not (live) physical
examination.
Conclusions
The 5-point photonumeric scale generated spans the
severity of the crow’s feet for which patients com-
monly seek correction. The scale is well stratified for
consistent rating.
References
1. American Academy of Ophthalmology. Botulinum toxin
(BOTOXs) for treating facial wrinkles: minimizing the signs
of aging [on-line]. Available at: http://medem.com/medlb/arti
cle_detaillb.cfm?article_ID=ZZZ7QNLURDD&sub_cat=2022.
Accessed July 5, 2007.
2. Fisher GJ. The pathophysiology of photoaging of the skin. Cutis
2005;75(Suppl 2):5–9.
3. American Academy of Dermatology. AgingSkinNet [on-line].
Available at: http://www.skincarephysicians.com/agingskinnet/
basicfacts.html. Accessed July 5, 2007.
4. Freiman A, Bird G, Metelitsa AI, et al. Cutaneous effects of
smoking. J Cutan Med Surg 2004;8:415–23.
5. Spiegel JH. Treatment of periorbital rhytides with botulinum
toxin type a maximizing safety and results. Arch Facial Plast Surg
2005;7:198–202.
6. Miller SR. Injectable fillers overview. June 2007. Available at:
http://www.yourplasticsurgeryguide.com/injectables-and-fillers/
injectable-fillers.htm. Accessed July 5, 2007.
7. Carruthers A, Carruthers J. Use of botulinum toxin A for facial
enhancement. In: Klein AW, editor. Tissue augmentation in clinical
practice. 2nd ed. New York, NY: Taylor & Francis; 2006.
p. 117–39.
S178 D E R M AT O L O G I C S U R G E RY
8. Kunzi-Rapp K, Dierickx CC, Cambier B, Drosner M. Minimally
invasive skin rejuvenation with erbium: YAG laser used in thermal
mode. Lasers Surg Med 2006;38:899–907.
9. Ronel DN, Spann M, Newman MI, Nolan WB. Skin resurfacing,
laser: erbium YAG. November 17, 2006 [on-line]. Available at:
http://www.emedicine.com/plastic/topic108.htm. Accessed
July 5, 2007.
10. Rubino C, Farace F, Dessy LA, et al. A prospective study of anti-
aging topical therapies using a quantitative method of assessment.
Plast Reconstr Surg 2005;115:1156–62.
11. Alexiades-Armenakas M. A quantitative and comprehensive
grading scale for rhytides, laxity, and photoaging. J Drugs
Dermatol 2006;5:808–9.
12. Kane MAC. Classification of crow’s feet patterns among Cauca-
sian women: the key to individualizing treatment. Plast Recon-
struct Surg 2003;112(Suppl):33S–9S.
13. Grove GL, Grove MJ, Leyden JJ. Optical profilometry: an
objective method for quantification of facial wrinkles. J Am
Acad Dermatol 1989;21(pt 2):631–7.
14. Hatzis J. The wrinkle and its measurementFa skin surface
profilometric method. Micron 2004;35:201–19.
15. Akazaki S, Nakagawa H, Kazama H, et al. Age-related
changes in skin wrinkles assessed by a novel three-
dimensional morphometric analysis. Br J Dermatol
2002;147:689–95.
16. Jacobi U, Chen M, Frankowski G, et al. In vivo determination of
skin surface topography using an optical 3D device. Skin Res
Technol 2004;10:207–14.
17. Heckmann M, Schön-Hupka G. Quantification of the efficacy of
botulinum toxin type A by digital image analysis. J Am Acad
Dermatol 2001;45:508–14.
18. Hund T, Ascher B, Rzany B. Reproducibility of two four-point
clinical severity scores for lateral canthal lines (crow’s feet).
Dermatol Surg 2006;32:1256–60.
19. Carruthers A, Carruthers J. Eyebrow height after botulinum
toxin type A to the Glabella. Dermatol Surg 2007;33:S26–31.
20. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater
reliability. Psychol Bull 1979;86:420–8.
21. Yaffee RA. Enhancement of reliability analysis: application
of intraclass correlations with SPSS/Windows v.8. New York:
Statistics and Social Science Group; 1998.
Address correspondence and reprint requests to: Alastair
Carruthers, MD, Suite 820-943 West Broadway, Vancou-
ver, British Columbia, Canada V5Z 4E1, or e-mail: alastair
@carruthers.net