CLINICAL SCIENCE

A New Method for Grading the Severity of Keratoconus

The Keratoconus Severity Score (KSS)
Timothy T. McMahon, OD,* Loretta Szczotka-Flynn, OD, MS, Joseph T. Barr, OD, MS,
Robert J. Anderson, PhD, Mary E. Slaughter, Jonathan H. Lass, MD,
Sudha K. Iyengar, PhD and the CLEK Study Group

Purpose: To define a new method for grading severity of

right and left eyes, respectively. Testretest analysis yielded k

statistics of 0.84 and 0.83 for right and left eyes, respectively.

keratoconus, the Keratoconus Severity Score (KSS).

Methods: A rationale for grading keratoconus severity was

developed using common clinical markers plus 2 corneal topographic
indices, creating a 0 to 5 severity score. An initial test set of 1012
eyes, including normal eyes, eyes with abnormal corneal and topographic findings but not keratoconus, and eyes with keratoconus
having a wide range of severity, was used to determine cutpoints for
the KSS. Validation set 1, comprising data from 128 eyes, was
assigned a KSS and compared with a clinicians ranking of severity
termed the gold standard to determine if the scale fairly represented
how a clinician would grade disease severity. k statistics, sensitivity,
and specificity were calculated. A program was developed to automate the determination of the score. This was tested against a manual
assignment of KSS in 2121 (validation set 2) eyes from the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study, as well
as normal eyes and abnormal eyes without keratoconus. Ten percent
of eyes underwent repeat manual assignment of KSS to determine the
variability of manual assignment of a score.

Results: From initial assessments, the KSS used 2 corneal topography indices: average corneal power and root mean square (RMS)
error for higher-order Zernike terms derived from the first corneal
surface wavefront. Clinical signs including Vogt striae, Fleischer
rings, and corneal scarring were also included. Last, a manual
interpretation of the map pattern was included. Validation set 1
yielded a k statistic of 0.904, with sensitivities ranging from 0.64 to
1.00 and specificities ranging from 0.93 to 0.98. The sensitivity and
specificity for determining nonkeratoconus from keratoconus were
both 1.00. Validation set 2 showed k statistics of 0.94 and 0.95 for

Received for publication November 8, 2005; revision received March 22,

2006; accepted March 25, 2006.
From the Department of Ophthalmology and Visual Sciences, University of
Illinois at Chicago, Chicago, IL; the Department of Ophthalmology, Case
Western Reserve University and University Hospitals of Cleveland,
Cleveland, OH; the Ohio State University College of Optometry,
Columbus, OH; the Division of Epidemiology and Biostatistics, School
of Public Health, University of Illinois at Chicago, Chicago, IL; and the
{Department of Epidemiology and Biostatistics, Case Western Reserve
University, Cleveland, OH.
Reprints: Timothy T. McMahon, OD, Department of Ophthalmology and
Visual Sciences University of Illinois at Chicago, Suite 3.164 (M/C 648),
1855 W. Taylor Street, Chicago, IL 60612 (e-mail: timomcma@uic.edu).
Copyright 2006 by Lippincott Williams & Wilkins

794

Conclusion: A simple and reliable grading system for keratoconus

was developed that can be largely automated. Such a grading scheme
could be useful in genetic studies for a complex trait such as
keratoconus requiring a quantitative measure of disease presence and
severity.
Key Words: keratoconus, severity, corneal topography, grading scale
(Cornea 2006;25:794800)

eratoconus is a bilateral noninflammatory corneal thinning

disorder leading to protrusion, distortion, and scarring of
the cornea.1 It is an uncommon disorder with widely variable
estimates of its annual incidence ranging from 50 to 230 per
100,000.2 Previous studies performed more than 20 years ago
estimate the prevalence to be 54 per 100,000.36 The origin of
keratoconus is unclear, although there is a growing body of
literature suggesting that in some cases keratoconus is
determined through heredity.730
Modern statistical genetics uses a variety of both modelbased and model-free methods to link genetic similarity to
phenotypic or clinical similarity. Phenotypes can simply be
classified as dichotomous variables (eg, affected vs. unaffected), or a more detailed characterization of a trait, such as
a severity index, can be used. Use of multiple discrete features
in an ordinal scale reduces the probability of misclassification
that is associated with simply classifying individuals as
affected versus unaffected to examine differences in the
genome of siblings and/or other relatives within a family. If
this type of technique is to be used in searching for a gene(s)
for keratoconus, a valid and easily applied severity scale is
needed. This requires that a scale be based on a variety of
phenotypic keratoconus features, broad enough to span the
range of disease severity from normal to severe, and be shown
as accurate with this categorization.
This report details the development of a new severity
scale, the Keratoconus Severity Score (KSS), based on slitlamp findings (including apical scarring), corneal topography
map characteristics, and 2 easily determined topographic
indices: average corneal power (ACP)31 and higher-order first
corneal surface wavefront root mean square (RMS) error
(HORMSE).3234 Such a severity scale could be applied to
a variety of circumstances, including genetic studies of
Cornea
Volume 25, Number 7, August 2006

Cornea
Volume 25, Number 7, August 2006

keratoconus, that require knowledge of disease severity

measured on an ordinal scale.

MATERIALS AND METHODS

Developing the Severity Score
Many indices derived from corneal topography were
initially evaluated as potential candidates for inclusion in the
new scale to meet the above-mentioned criteria. Simulated
indices were computed and incorporated in the Ohio State
Corneal Topography Tool (OSUCTT).35 These indices are
simulated because they were initially derived from descriptions
in the published literature and compared with the results
provided by the proprietary instrument using them. For example,
ACP was developed for the TMS-1 instrument (Tomey
Technologies, Nagoya, Japan).31 In our previous studies, when
ACP was compared with the simulated ACP, there were
discrepancies that could be explained only by manufacturer
modifications to the published formulas (unpublished data). In
fact, this was the rule rather than the exception across instrument
manufacturers for other simulated indices as well. These
discrepancies required reverse engineering to improve the
correlations between the native instrument output and that
derived from the OSUCTT. Hence, for consistency across
platforms, we used simulated indices in this study.
The first set of analyses to identify candidate topographic indices examined the correlation for several pairs of
indices. This analysis grouped 17 indices into 4 major groups:
those associated with corneal power, corneal asymmetry,
corneal irregularity, and corneal cylinder (orthogonal astigmatism). Among the corneal power measures, the correlations
were very high. ACP was chosen because the simulations were
closest to the native machine output. Corneal asymmetry and
irregularity essentially represent higher-order optical aberrations, so we selected the third-order and higher RMS error
(HORMSE; through the 27th term) to describe these features
collectively. HORMSE was derived from raw corneal
topography data by VOL-CT software (version 6.58; Sarver
and Associates, Carbondale, IL). Corneal astigmatism was not
used in this scale. In previous studies (unpublished data), the
cylinder was not a very powerful diagnostic tool either in
identifying keratoconus or in tracking its severity.
In addition to these 2 topographic indices, the KSS also
used an analysis of topographic patterns. One observer
(T.T.M.) subjectively classified the topographic patterns using
an axial algorithm-displayed map as either normal, atypical
(but not keratoconus), or whether it exhibited an isolated area
of steepening characteristic for keratoconus. Finally, the KSS
used clinical assessments of keratoconic slit-lamp signs,
specifically the presence or absence of Fleischer rings, Vogt
striae, and corneal scarring characteristic for keratoconus. The
scaling of scarring for KSS followed the protocol used for
gestalt scarring in the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study.36
Gestalt scarring is a measure of the total scarring
observed in the central cornea in CLEK Study corneal
photographs. In a previous study, masked CLEK Photography
Reading Center readers assessed the number, size, density, and
q 2006 Lippincott Williams & Wilkins

Severity Paper

shape of corneal opacities in corneas of keratoconus patients.36

In the Gestalt scarring assessment, the reader takes into account how close the scarring as a whole is to the line of sight
and how large and dense is the scarring to estimate the overall
(gestalt) scarring. Table 1 defines the gestalt scarring scale.
Intraclass correlation coefficients for readers reading the same
slides of corneal scarring (masked) on a repeated basis indicate
very good reliability.37 The testretest reliability for gestalt
grading for random rereads for all readers for each year from
1998 to 2004 were 0.77, 0.49, 0.80, 0.71, 0.54, 0.72, and 0.61,
respectively. Over this 6-year period, the testretest reliability
was 0.66, which is typically considered very good (unpublished data).
A scale suitable for segregation analysis requires a range
of quantitative classifications from normal to severely affected
cases. Our scale includes values for normal eyes, keratoconus
suspects, and mild, moderate, and severe keratoconus. Unusual corneas not caused by keratoconus referred to as atypical
corneas were also included. Atypical eyes consisted of contact
lensinduced warpage, penetrating keratoplasty, myopic
refractive surgery, and corneal scarring from disease or trauma
not associated with keratoconus. This resulted in a 5-point scale,
with 0 being normal and 5 indicating severe disease.
Testing of data consisted of 3 steps: evaluating a test
dataset and validating the scale results using 2 validation sets.

Test Dataset
An initial test set of 1012 eyes was assembled to
determine the combination of data ranges for each classification criterion. The test set included subjects that clinically
could be classified into each of these categories. The set
included 130 normal eyes, 41 atypical eyes, 7 keratoconussuspect eyes, and 834 eyes with keratoconus. Keratoconus
suspects had corneal topography patterns suspicious for the
disease but no slit-lamp or other clinical findings. Using 95%
confidence intervals (CIs), possible demarcation points were
determined for HORMSE for each level. The cutpoints for
ACP were determined through clinical experience and 95% CI
and were defined as less than 52.00 D = mild, 52.00 to 56.00
D = moderate, and more than 56.00 D = severe.

Validation Set 1
A set of 128 right eyes, referred to as validation set 1,
was subjected to ranking and compared with a clinicians

TABLE 1. Descriptors for Overall (Gestalt) Scarring (0.04.0

in 0.5 Steps)
Grade 1.0
Grade 2.0
Grade 3.0
Grade 4.0

Trace and not on LOS,

,1.5 mm total size
Easily noticeable and approaching LOS,
1.52.5 mm total size
Dense but translucent and impinging on LOS,
total size 2.5 mm or greater
Opaque and on LOS, size 2.5 mm or greater

Increased grade for density, size, number, and location near or on the line of
sight (LOS).

795

McMahon et al

grading of severity used in clinical practice, which we refer to

as the gold standard. The gold standard was determined in
validation set 1 by one of the authors (L.S.-F.) using clinical
chart data including slit-lamp findings, best-corrected visual
acuity, color corneal topography maps from the Keratron or
Humphrey Atlas systems, and keratometry. The 2 topographically derived indices, ACP and HORMSE, were not available
for the gold standard grading. One hundred eyes from CLEK
Study subjects were examined at the Department of Ophthalmology at the Case Western Reserve University CLEK
site. Also, 10 normals, 10 atypical normals, and 8 suspect
cases (not in the CLEK Study) were also provided by the Case
Western University, Department of Ophthalmology clinic.
Because there is no uniformly accepted severity classification
scheme for keratoconus using a combination of objective and
subjective assessments, the gold standard grading criterion
was independently developed by the examiner (L.S.-F.) using
clinical experience in combination with published recommendations.3842 The gold standard grading system is described in Table 2. These gold standard grades were compared
with the KSS determined for this same group of patients at the
Corneal Topography Reading Center (CTRC) at the University
of Illinois at Chicago. Corneal topography data, computed
indices, and corneal slit-lamp signs including the corneal
scarring grade provided by the CLEK Study database were
used by the CTRC to establish a KSS.
Cohen k statistics were calculated. Also, 2 3 2
contingency tables were constructed to calculate the sensitivity
and specificity for determining the screening capability of
the model in differentiating keratoconus from normals and
atypical normals and identifying the different levels of
keratoconus severity.
To automate the classification of KSS, a cascading
classification algorithm was written using the SAS statistical
software package (version 9.1). The algorithm used the
structure and criteria of the KSS (Table 3). The decision
process within the algorithm flowed down the 6 grades,
assigning the largest KSS score possible with all required
criteria within a grade satisfied. A KSS score was not assigned
for an eye that underwent a corneal transplant or if any data
needed to assign a score were missing.

Validation Set 2
To validate the assignment of a KSS score, a second
validation set (validation set 2) was assembled. This set
consisted of a cohort of CLEK Study subjects with
keratoconus plus keratoconus suspects, normals, and atypical
normals.43 For those CLEK eyes with no missing data,
a manual KSS was determined using a combination of the
CLEK data, topographical scans, and clinical expertise from
the grader (T.T.M.). The agreement between the algorithm and
the manual KSS was evaluated with k statistics. Also, to assess
testretest reliability of the observer (T.T.M.) in manually
assigning a KSS, a random sample of approximately 10% of
validation set 2 eyes (138 total) was assessed again in a
masked manner, both manually and with a second pass (retest)
of the KSS scoring algorithm. k statistics, evaluating the

796

Cornea
Volume 25, Number 7, August 2006

TABLE 2. Gold Standard Grading Scheme

Normal
Regular axial topography pattern (round, oval, symmetric bow tie, etc)
Normal slit-lamp exam
Spectacle corrected acuity P55 letters at 4 m on Log Mar chart (with
no other ocular pathology)
Atypical normal
Unusual axial topography explained by slit-lamp exam or history
(contact lens warpage, corneal scars not typical of keratoconus, history
of refractive surgery)
Normal or diminished spectacle acuity
Keratoconus suspect
Suspicious axial topography for keratoconus (isolated area of steepening,
central steepening .48 D)
Normal slit-lamp exam
Spectacle corrected acuity P55 letters at 4 m on Log Mar chart (with no
other ocular pathology)
Mild keratoconus
Axial topography consistent with keratoconus
Flat keratometry reading ,51.00 D
Fleischer ring or Vogt striae
No corneal scarring
Reduced spectacle acuity (,55 letters at 4 m on Log Mar chart) (with no
other ocular pathology)
Moderate keratoconus
Axial topography consistent with keratoconus
Flat keratometry reading between 51.25 and 56.00 D or astigmatism P8 D
Fleischer ring or Vogt striae
May have corneal scarring up to and including CLEK grade 3.0 (any
scarring up to well-defined stromal scarring consistent with keratoconus)
Reduced spectacle acuity ( 645 letters at 4 m on Log Mar chart)
(with no other ocular pathology)
Severe keratoconus
Axial topography consistent with keratoconus with marked areas
of steepening
Flat keratometry reading .56.01 D
Fleischer ring or Vogt striae
May have corneal scarring up to and including CLEK grade 4.0
(any scarring up to a dense/opaque stromal scar consistent
with keratoconus)
Reduced spectacle acuity (,30 letters at 4 m on Log Mar chart)
(with no other ocular pathology)

agreement between the manual observerassigned and the

calculated KSS, were determined. These validations were
performed to assess variability of the observer and algorithm
to accurately assign a KSS score to the CLEK cohort.

RESULTS
Table 4 shows the mean, range, SD, and 95% CIs for
ACP and HORMSE for each level for the initial test set. The
KSS scale is defined in Table 3. It is derived largely from the
data found in Table 2 and the application of judicious clinical
opinion. The strategy to rank an eye is dependent on the
worst feature of any of the data types in the scale algorithm.
The final 5 features assessed were slit-lamp signs, topography
q 2006 Lippincott Williams & Wilkins

Cornea
Volume 25, Number 7, August 2006

TABLE 3. Keratoconus Severity Score Ranking Scheme

0

Unaffectednormal topography
Required features:
No corneal scarring consistent with keratoconus
No slit-lamp signs for keratoconus
Typical axial pattern
Average corneal power (ACP) #47.75 D
Higher-order RMS error #0.65
Unaffectedatypical topography
Required features:
No corneal scarring consistent with keratoconus
No slit-lamp signs for keratoconus
Atypical axial pattern
Irregular pattern
or
Asymmetric superior bowtie
or
Asymmetric inferior bowtie
or
Inferior or superior steepening no more than
3.00 D steeper than ACP
ACP #48.00 D
Higher-order RMS error #1.00
Suspect topography
Required features:
No corneal scarring consistent with keratoconus
No slit-lamp signs for keratoconus
Axial pattern with isolated area of steepening
Inferior steep pattern
or
Superior steep pattern
or
Central steep pattern
Additional features:
ACP #49.00 D
or
Higher-order RMS error .1.00, #1.50
Affectedmild disease
Required features:
Axial pattern consistent with KCN
May have positive slit-lamp signs
No corneal scarring consistent with keratoconus
Additional features:
ACP #52.00 D
or
Higher-order RMS error .1.50, #3.50
Affectedmoderate disease
Required features:
Axial pattern consistent with KCN
Must have positive slit-lamp signs
Additional features:
ACP .52.00 D, #56.00 D
or
Higher-order RMS error .3.50, #5.75
or
Corneal scarring and overall CLEK grade up to 3.0
Affectedsevere disease

q 2006 Lippincott Williams & Wilkins

Severity Paper

TABLE 3. (Continued)
Required features:
Axial pattern consistent with KCN
Must have positive slit-lamp signs
Additional features:
ACP .56.00 D
or
Higher-order RMS error .5.75
or
Corneal scarring CLEK grade 3.5 or greater overall
Rules: The decision process flows down each grade. For grades 01, all of the
parameters in a category must be met. For all grades, the required features must be met.
The worst of the additional features is then assessed, with the worst of the features
carrying the greater weight (as long as the required features are met).

pattern, corneal scarring, ACP, and HORMSE. For normals

and atypical normals, all of the features of the category must
have been met for an eye to assume the KSS score for that
category. At the suspect level in the scale, the decision-making
was bifurcated. To be placed in the suspect category, an eye
must have had no scarring, no other slit-lamp findings for
keratoconus, and have had an axial topography pattern with an
isolated area of steepening typical for keratoconus. In addition,
the worse of ACP or HORMSE defined the KSS category. This
bifurcated decision tree extended to all the higher levels, with
changing criteria as the severity increased.
k statistics were calculated for the components of the
model to determine the relative value of each component
(Table 5). As can be seen, the addition of each successive
component to the model increases the k statistic, thus enhancing the fit to the gold standard evaluation. Table 6 defines the
sensitivity and specificity of KSS to segregate eyes into the
proper score. This analysis was set up for each grade level
computed for the KSS and compared with the gold standard by
using validation set 1.
The eyes in validation set 2, including the CLEK Study
cohort, were assigned a KSS by using the algorithm in Table 5.
These eyes encompassed the entire range of the scale; 57
normal eyes, 8 atypical eyes, 49 keratoconus suspects, 927
eyes with mild keratoconus, 682 with moderate disease, and
398 with severe keratoconus. To determine the consistency
with which the algorithm assigned KSS compared with a
manually assigned KSS, weighted k statistics were calculated
using CicchettiAllison weights.44 For right eyes, the weighted
k was 0.94 (95% CI: 0.920.96); for left eyes, it was 0.95
(95% CI: 0.940.97). Both of these statistics indicate very high
agreement and give confidence to the use of an automated
approach for assigning a KSS score.
A random sample of validation set 2 comprising approximately 10% of the overall sample was used to assess the
testretest reliability of assigning a severity score. There were
69 right eyes and 69 left eyes in the sample (eyes with
a corneal transplant were excluded). These maps were rescored
weeks after their initial score assignment. Eyes used in this
sample had initial KSS ranging from 1 through 5. Weighted k
statistics were 0.84 (95% CI: 0.730.95) for right eyes and
0.83 (95% CI: 0.710.94) for left eyes.

797

Cornea
Volume 25, Number 7, August 2006

McMahon et al

TABLE 4. Summary Statistics for 2 Keratoconus Measures, by KSS Level, Obtained from Test Set
ACP
Mean
Min
Max
SD
95% CI
HORMSE
Mean
Min
Max
SD
95% CI

43.82
41.10
47.66
1.30
41.2246.42

44.10
40.00
47.01
1.60
40.9044.30

44.57
40.86
46.83
2.24
40.0949.05

47.10
39.47
51.97
2.74
41.6252.58

53.89
52.01
55.98
1.11
51.6756.11

62.81
56.08
90.38
6.52
49.7775.85

0.42
0.23
0.75
0.07
0.280.56

0.60
0.40
1.0
0.14
0.320.88

0.70
0.51
0.96
1.61
03.93

1.98
0.26
9.48
1.19
04.36

3.14
0.89
8.00
1.30
0.545.74

5.03
1.11
55.66
4.27
013.57

Scores: 0, normal; 1, atypical normal; 2, keratoconus suspect; 3, keratoconusmild; 4, keratoconusmoderate; 5, keratoconussevere.

ACP, average corneal power; HORMSE, higher-order root mean square (Wavefront) error.

DISCUSSION
The detection of keratoconus has received a great deal of
attention in the past 15 years, concomitant with the rise in
popularity of refractive surgery. The development of Placido
disk-based videokeratography was stimulated largely by the
desire to screen out patients with keratoconus from the group
of prospective refractive surgery candidates.45 Corneal topography has proved valuable in identifying cases of subtle or
forme fruste keratoconus.4549 Several analytical, topographybased screening tools have been developed to detect eyes with
signs of keratoconus.31,5052 These tools have limited use as
screening tools in most cases and suffer significant shortcomings in actually tracking the severity of keratoconus as it
progresses.53,54
There are a few techniques that have been developed for
tracking disease severity in keratoconus. Smolek and Klyce
have developed a Keratoconus Severity Index (KSI) using
previously developed expert systems and artificial intelligence.31 This system is proprietary to 1 instrument and to use it
with corneal topography data from other instruments would
not be prudent until appropriate model training and validation
has been accomplished. To date, this has not been available.
Avitabile et al55,56 have used an ultrasound biomicroscope (UBM) technique for grading and tracking disease
severity in keratoconus that compares favorably with corneal
topographybased KSI readings. However, although less

TABLE 5. Agreement (Computed With Cohen k Statistic)

Between KSS and the Gold Standard Evaluation
Weighted k

95% CI

0.695
0.854
0.762
0.863

0.5980.792
0.7990.910
0.6780.847
0.8070.920

0.904

0.8620.946

proprietary, UBMs are uncommon instruments, and the transfer of the authors technique to the clinical environment has
not been reported.
Li et al38 have developed an index, the KISA%, to grade
the presence or absence of keratoconus. Although this index
has the potential to define disease severity in keratoconus, the
developers have described its role only in defining normals,
keratoconus suspects, and those with the disease. The KISA%
index has been used to monitor changes in normal eyes of
unilateral keratoconus patients38 and in genetic screening,
where KISA% was used to distinguish keratoconus from
normal individuals.11,57
Rabinowitz58 has described a classification scheme of
3 distinct categories: keratoconus, early keratoconus, and
keratoconus suspect. In his most advanced categorization
(keratoconus), disease can be detected by slit-lamp evaluation
and an asymmetric bowtie/skewed radial axis pattern
(AB/SRAX) on videokeratography. In early keratoconus, no
slit-lamp findings of disease are found, but scissoring of the
retroilluminated reflex and an AB/SRAX pattern are present.
In keratoconus suspects, no clinical signs of keratoconus on
either slit-lamp evaluation or retroillumination assessment are
found, but there is an AB/SRAX pattern. In our experience,
there are frequent occasions (eg, an isolated area of inferior
steepening) where eyes with keratoconus did not have
a definite AB/SRAX pattern.

TABLE 6. Sensitivity and Specificity for Screening for

Assignment of Proper Grade Level
Grade Description (Grade Number)

Indices only
Indices and manual read
Indices and scarring
Indices, manual read, and scarring
Indices, manual read, scarring,
and slit-lamp signs

798

Normal (0) vs. atypical normal (1)

Atypical normal (1) vs. keratoconus suspect (2)
Keratoconus suspect (2) vs. mild keratoconus (3)
Mild keratoconus (3) vs. moderate keratoconus (4)
Moderate keratoconus (4) vs. severe keratoconus (5)

Sensitivity Specificity
1.00
1.00
1.00
0.90
0.64

0.95
1.00
1.00
0.93
0.98

q 2006 Lippincott Williams & Wilkins

Cornea
Volume 25, Number 7, August 2006

In other classification schemes, the shape of the conenipple (round), oval, or globus has been used to classify
keratoconus, although these do not lend themselves to grading
severity. Last, Hom59 classified keratoconus into 4 stages, with
stage 1 having such early disease that spectacles are the first
form of treatment, and stage 4 has corneal steepening .55 D,
apical opacities, and Munson sign.
As previously mentioned, there is a growing body of
evidence that keratoconus has a hereditary component or is
fully an inherited disease. Despite many studies, there is little
concordance of results as to the location of a gene or genes for
keratoconus.11,12,14,20,21,60,61 In fact, it is remarkable how
dissimilar the regions of the genome are that have been suggested. One potential explanation would be that the composite
keratoconus trait is a complex genetic trait. A good model to
discover complex genetic traits includes a reliable means of
defining disease severity. Other than a few proprietary topography instruments (based solely on topography data), no such
method currently exists.
We present a severity rating scale, based on common
clinical signs and easily obtained corneal topography indices,
that corresponds well with clinician grading of severity, has
very good reproducibility, and has a strong ability to separate
normal corneas and those with abnormal but nonkeratoconic
topographic features (such as refractive surgery or trauma)
from those with keratoconus. This method also is not tied to
any 1 particular topography instrument. It requires a skilled
observer to detect clinical slit-lamp signs of keratoconus and
an observer competent in reviewing topography maps for some
segregation into the proper severity score. However, objective
reads of masked data, computer-derived indices, and ultimately, computer-driven classification assignments are also
incorporated, which make this scale beneficial for use in
multicenter studies relying on a quantitative classification
scheme of keratoconus.

REFERENCES
1. Krachmer JH, Feder RS, Belin MW. Keratoconus and related noninflammatory corneal thinning disorders. Surv Ophthalmol. 1984;28:
293322.
2. Rabinowitz YS. Keratoconus. Surv Ophthalmol. 1998;42:297319.
3. Hoffstetter H. A keratoscopic survery of 13,395 eyes. Am J Optom Acad
Optom. 1959;36:311.
4. Duke-Elder S, Leigh AG. Diseases of the cornea. In: Duke-Elder S, ed.
System of Ophthalmology (vol. III), London: Henry Kimpton; 1965:976.
5. Franceschetti A. Keratoconus. In: King JH, McTigue JW, eds. The Cornea
World Congress. Washington, DC: Butterworths; 1965:511518.
6. Kennedy RH, Bourne WM, Dyer JA. A 48-year clinical and epidemiologic study of keratoconus. Am J Ophthalmol. 1986;101:267273.
7. Bisceglia L, Ciaschetti M, De Bonis P, et al. VSX1 mutational analysis in
a series of Italian patients affected by keratoconus: detection of a novel
mutation. Invest Ophthalmol Vis Sci. 2005;46:3945.
8. Rivera L, Mendoza SD. Keratoconus in Puerto Rico. Bol Asoc Med P R.
2004;96:9496.
9. Ha NT, Nakayasu K, Murakami A, et al. Microarray analysis identified
differentially expressed genes in keratocytes from keratoconus patients.
Curr Eye Res. 2004;28:373379.
10. Javadi MA, Rafeei AB, Kamalian N, et al. Concomitant keratoconus and
macular corneal dystrophy. Cornea. 2004;23:508512.
11. Levy D, Hutchings H, Rouland JF, et al. Videokeratographic anomalies in
familial keratoconus. Ophthalmology. 2004;111:867874.

q 2006 Lippincott Williams & Wilkins

Severity Paper

12. Brancati F, Valente EM, Sarkozy A, et al. A locus for autosomal dominant
keratoconus maps to human chromosome 3p14-q13. J Med Genet. 2004;
41:188192.
13. Rabinowitz YS. The genetics of keratoconus. Ophthalmol Clin North Am.
2003;16:607620. vii.
14. Hughes AE, Dash DP, Jackson AJ, et al. Familial keratoconus with
cataract: linkage to the long arm of chromosome 15 and exclusion of
candidate genes. Invest Ophthalmol Vis Sci. 2003;44:50635066.
15. Bochert A, Berlau J, Koczan D, et al. [Gene expression in keratoconus.
Initial results using DNA microarrays]. Ophthalmologe. 2003;100:545549.
16. Nielsen K, Birkenkamp-Demtroder K, Ehlers N, et al. Identification of
differentially expressed genes in keratoconus epithelium analyzed on
microarrays. Invest Ophthalmol Vis Sci. 2003;44:24662476.
17. Owens H, Gamble G. A profile of keratoconus in New Zealand. Cornea.
2003;22:122125.
18. Igarashi S, Makita Y, Hikichi T, et al. Association of keratoconus and
Avellino corneal dystrophy. Br J Ophthalmol. 2003;87:367368.
19. Tachibana M, Okamoto M, Sakamoto M, et al. Hereditary keratoconuslike keratopathy in Japanese wild mice mapped to mouse chromosome 13.
Mamm Genome. 2002;13:692695.
20. Tyynismaa H, Sistonen P, Tuupanen S, et al. A locus for autosomal
dominant keratoconus: linkage to 16q22.3-q23.1 in Finnish families.
Invest Ophthalmol Vis Sci. 2002;43:31603164.
21. Heon E, Greenberg A, Kopp KK, et al. VSX1: a gene for posterior
polymorphous dystrophy and keratoconus. Hum Mol Genet. 2002;11:
10291036.
22. Fullerton J, Paprocki P, Foote S, et al. Identity-by-descent approach to
gene localisation in eight individuals affected by keratoconus from northwest Tasmania, Australia. Hum Genet. 2002;110:462470.
23. Maruyama Y, Li Y, Zhang Y, et al. Mapping of Sp1 regulation sites in the
promoter of the human alpha1-proteinase inhibitor gene. J Cell Biochem.
2002;85:482489.
24. Edwards M, McGhee CN, Dean S. The genetics of keratoconus. Clin
Experiment Ophthalmol. 2001;29:345351.
25. Morrison DA, Rosser EM, Claoue C. Keratoconus associated with
a chromosome 7,11 translocation. Eye. 2001;15:556557.
26. Maruyama Y, Wang X, Li Y, et al. Involvement of Sp1 elements in the
promoter activity of genes affected in keratoconus. Invest Ophthalmol Vis
Sci. 2001;42:19801985.
27. Heaven CJ, Lalloo F, McHale E. Keratoconus associated with
chromosome 13 ring abnormality. Br J Ophthalmol. 2000;84:1079.
28. Bechara SJ, Waring GO 3rd, Insler MS. Keratoconus in two pairs of
identical twins. Cornea. 1996;15:9093.
29. McMahon TT, Shin JA, Newlin A, et al. Discordance for keratoconus in
two pairs of monozygotic twins. Cornea. 1999;18:444451.
30. Rabinowitz YS, Maumenee IH, Lundergan MK, et al. Molecular genetic
analysis in autosomal dominant keratoconus. Cornea. 1992;11:302308.
31. Smolek MK, Klyce SD. Current keratoconus detection methods compared
with a neural network approach. Invest Ophthalmol Vis Sci. 1997;38:
22902299.
32. Cheng X, Himebaugh NL, Kollbaum PS, et al. Test-retest reliability of
clinical Shack-Hartmann measurements. Invest Ophthalmol Vis Sci. 2004;
45:351360.
33. Cheng X, Himebaugh NL, Kollbaum PS, et al. Validation of a clinical
Shack-Hartmann aberrometer. Optom Vis Sci. 2003;80:587595.
34. Thibos LN, Hong X. Clinical applications of the Shack-Hartmann
aberrometer. Optom Vis Sci. 1999;76:817825.
35. Mahmoud AM, Roberts C, Henderick EE. The Ohio State University
Corneal Topography Tool [ARVO Abstract]. Invest Ophthalmol Vis Sci.
2000;41:S677. Abstract nr 3599B697.
36. Barr JT, Wilson BS, Gordon MO, et al. Estimation of the incidence and
factors predictive of corneal scarring in the Collaborative Longitudinal
Evaluation of Keratoconus (CLEK) Study. Cornea. 2006;25:1625.
37. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater
reliability. Psychol Bull. 1979;86:420428.
38. Li X, Rabinowitz YS, Rasheed K, et al. Longitudinal study of the normal
eyes in unilateral keratoconus patients. Ophthalmology. 2004;111:440446.
39. Mandell RB. Contemporary management of keratoconus. ICLC. 1997;24:
4358.
40. Rabinowitz YS, Garbus J, McDonnell PJ. Computer-assisted corneal topography in family members of patients with keratoconus. Arch Ophthalmol.
1990;108:365371.

799

McMahon et al

41. Vitale S. CLEK study reports on the quality of life. Am J Ophthalmol.

2004;138:637638.
42. Zadnik K, Barr JT, Gordon MO, et al. Biomicroscopic signs and disease
severity in keratoconus. Collaborative Longitudinal Evaluation of
Keratoconus (CLEK) Study Group. Cornea. 1996;15:139146.
43. Zadnik K, Barr JT, Edrington TB, et al. Baseline findings in the
Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study.
Invest Ophthalmol Vis Sci. 1998;39:25372546.
44. Cicchetti DV, Allison T. A new procedure for assessing reliability of
scoring EEG sleep recordings. Am J EEG Tech. 1971;11:101109.
45. Wilson SE, Klyce SD. Screening for corneal topographic abnormalities
before refractive surgery. Ophthalmology. 1994;101:147152.
46. Maeda N, Klyce SD, Tano Y. Detection and classification of mild irregular
astigmatism in patients with good visual acuity. Surv Ophthalmol. 1998;
43:5358.
47. Maguire LJ, Bourne WM. Corneal topography of early keratoconus. Am J
Ophthalmol. 1989;108:107112.
48. Maguire LJ, Lowry JC. Identifying progression of subclinical keratoconus
by serial topography analysis. Am J Ophthalmol. 1991;112:4145.
49. Rabinowitz YS, Rasheed K, Yang H, et al. Accuracy of ultrasonic
pachymetry and videokeratography in detecting keratoconus. J Cataract
Refract Surg. 1998;24:196201.
50. Maeda N, Klyce SD, Smolek MK, et al. Automated keratoconus screening
with corneal topography analysis. Invest Ophthalmol Vis Sci. 1994;35:
27492757.
51. Smolek MK, Klyce SD. Screening of prior refractive surgery by a waveletbased neural network. J Cataract Refract Surg. 2001;27:19261931.

800

Cornea
Volume 25, Number 7, August 2006

52. Klyce SD, Wilson SE, Kaufman HE. Corneal topography comes of age.
Refract Corneal Surg. 1989;5:359361.
53. Klyce SD, Smolek MK, Maeda N. Keratoconus detection with the KISA%
methodanother view. J Cataract Refract Surg. 2000;26:472474.
54. Rabinowitz YS. Corneal topography. Curr Opin Ophthalmol. 1995;6:
5762.
55. Avitabile T, Franco L, Ortisi E, et al. Keratoconus staging: a computerassisted ultrabiomicroscopic method compared with videokeratographic
analysis. Cornea. 2004;23:655660.
56. Avitabile T, Marano F, Castiglione F, et al. Keratoconus staging
with ultrasound biomicroscopy. Ophthalmologica. 1998;212 (Suppl 1):
1012.
57. Wang Y, Rabinowitz YS, Rotter JI, et al. Genetic epidemiological study of
keratoconus: evidence for major gene determination. Am J Med Genet.
2000;93:403409.
58. Rabinowitz YS. Corneal topography. In: Bennett ES, Weissman BA, eds.
Clinical Contact Lens Practice. Philadelphia, PA: Lippincott, Wlliams
and Wilkins; 2005:215232.
59. Hom M., Bruce AS. Manual of Contact Lens Prescribing and Fitting.
London: Butterworth-Heineman, 2006:503544.
60. Hutchings H, Ginisty H, Le Gallo M, et al. Identification of a new
locus for isolated familial keratoconus at 2p24. J Med Genet. 2005;42:
8894.
61. Tang YG, Rabinowitz YS, Taylor KD, et al. Genomewide linkage
scan in a multigeneration Caucasian pedigree identifies a novel locus
for keratoconus on chromosome 5q14.3-q21.1. Genet Med. 2005;7:
397405.

q 2006 Lippincott Williams & Wilkins