Comparative Corneal Topography and
Refractive Variables in Monozygotic and
Dizygotic Twins
SHAILAJA VALLURI, MD, JEFFREY B. MINKOVITZ, MD, KORAY BUDAK, MD,
LYDIA R. ESSARY, MD, REBECCA S. WALKER, MD, EKKTET CHANSUE, MD,
G. M. CABRERA, MD, DOUGLAS D. KOCH, MD, AND JAY S. PEPOSE, MD, PHD
● PURPOSE: To investigate the role of heredity in deter-
mining corneal shape, axial length, and overall refractive
error.
● METHODS: Twenty monozygotic and 19 dizygotic twin
pairs, age 12 to 73 years, were enrolled in the study.
Zygosity was determined by physical similarity and by
responses to questions adapted from surveys. Two twin
pairs were excluded because of undetermined zygosity
and one pair because of keratoconus (both siblings).
Refractive error was determined by an automated refrac-
tor. Manifest refraction was also recorded, as well as
cycloplegic refraction in subjects under age 18 years.
Corneal topography data and manual keratometer read-
ings were also obtained. Axial lengths were determined
by A-scan ultrasound. Data were analyzed by Student t
tests only in the right eye. Left-eye data were comparable
for all variables.
● RESULTS: Mean intrapair difference in refractive error
(spherical equivalent) was less for monozygotic than for
dizygotic twins (RE: 0.41 vs 1.53; P ! .001). Mean
intrapair difference in axial length was less for monozy-
gotic twins (RE: 0.39 vs 0.76 mm; P ! .031). Corneal
topography data (power and meridian) in all zones (3, 5,
and 7 mm) also showed smaller mean differences among
monozygotic pairs than dizygotic, but the difference was
statistically significant only for the 5-mm zone. In addi-
tion, most Holladay Diagnostic Summary variables that
were studied did not show any statistically significant
differences.
Accepted for publication Aug 27, 1998.
From the Department of Ophthalmology and Visual Sciences, Wash-
ington University School of Medicine, St Louis, Missouri (Drs Valluri,
Minkovitz, Essary, Walker, Chansue, and Pepose); and Cullen Eye
Institute, Baylor College of Medicine, Houston, Texas (Drs Budak,
Cabrera, and Koch).
Supported in part by an unrestricted grant from Research to Prevent
Blindness, Inc, New York, New York, and core grant EY02687 from the
National Institutes of Health, Bethesda, Maryland.
Reprint requests to Jay S. Pepose, MD, PhD, Department of Ophthal-
mology and Visual Sciences, Washington University School of Medicine,
660 S Euclid, Box 8096, St Louis, MO 63110.
158 © 1999 BY ELSEVIER SCIENCE INC. ALL
● CONCLUSIONS: Axial length and overall refractive
error have a significant genetic basis. Corneal topography
data appear to have other overriding determining factors
for several of the variables studied. (Am J Ophthalmol
1999;127:158 –163. © 1999 by Elsevier Science Inc.
All rights reserved.)
B
EFORE THE 20TH CENTURY, AMETROPIA WAS CONSID-
ered to be a pathologic state caused by the alteration
of axial length, mechanically induced by traction or
pressure.1 It is now understood that several components
contribute to refractive error, including axial length, lens
power, corneal power, and anterior chamber depth. Whether
ametropia is genetically or environmentally determined
has been the subject of much debate. Raviola and Wiesel2
found that suturing the eyelids of macaque monkeys before
completion of growth promoted axial myopia, but only if
diffuse light was allowed to reach the sutured lids. In
contrast, early studies by Sorsby and associates3 on twins
yielded much data supporting a genetic basis for refractive
error.
The present study of refractive and corneal topographic
variables in monozygotic and dizygotic twins was under-
taken to determine the genetic and environmental influ-
ences on refractive components by means of modern
measurement techniques, including computerized corneal
videokeratography and ultrasonography.
METHODS
MONOZYGOTIC AND DIZYGOTIC TWIN PAIRS, AGE 12 TO 80
years, were recruited for the study through television,
radio, and bulletin board advertising. Twins were screened
by telephone for any history of previous eye disease or
surgery. This study was approved by the Washington
University human studies committee. Soft contact lens
wearers were asked to refrain from all lens wear for a
minimum of 2 days before examination, and wearers of
RIGHTS RESERVED. 0002-9394/99/$20.00
PII S0002-9394(98)00319-5
hard or rigid gas-permeable lenses were asked to discon-
tinue contact lens use for 2 weeks before examination. All
contact lens wearers, with the exception of one twin pair,
subsequently declined to participate. Twin pairs were
examined together to facilitate zygosity determination.
Participants filled out questionnaires detailing ocular his-
tory. Information aimed at determining zygosity was also
obtained with the use of questions adapted from surveys by
Sarna and associates.4 Examination of each twin included
the subjective assessment of similarity, as well as the
recording of eye and hair color, height, and weight. Visual
acuity was measured with the standardized Early Treat-
ment Diabetic Retinopathy Study chart. Refraction was
performed with an automated refractor (Nidek Inc, Palo
Alto, California) followed by a manifest refraction. Cyclo-
plegic autorefraction with cyclopentolate 1% was per-
formed in all participants under age 18 years. Keratometry
was measured with a keratometer (Bausch and Lomb Inc,
Buffalo, New York). Slit-lamp examination was performed
to rule out any ocular disease. Finally, axial length mea-
surements were made with Storz A-scan ultrasound (Storz
Ophthalmics Inc, St Louis, Missouri).
In addition, computed corneal topographic analysis was
performed (EyeSys Corneal Analysis System, Version 3.2;
EyeSys Technologies, Inc, Houston, Texas).5 We com-
pared the intrapair differences in magnitude and meridian
of astigmatism at the 3-, 5-, and 7-mm zones. The axial,
instantaneous, refractive, and profile difference maps were
obtained for each eye and evaluated by means of a scale of
0.5-D intervals. Patterns were classified by a masked
observer (K.B.) with respect to middioptric green value by
means of the classification reported in Budak and associ-
ates.6
Intrapair differences for the following variables from
Holladay Diagnostic Summary7 were also evaluated. The
steep refractive power is the highest refractive power in
any single meridian of the cornea within the 3-mm pupil
zone. The flat refractive power is the lowest refractive
power in any single meridian of the cornea within the
3-mm pupil zone. The effective refractive power is the
refractive power of the corneal surface within the 3-mm
pupil zone. Total astigmatism is the difference between the
flat and steep refractive powers within the 3-mm pupil
zone; intrapair differences for total astigmatism were eval-
uated by the vector analysis method used by Holladay and
associates.8 Irregular astigmatism is the difference between
total astigmatism and regular astigmatism (amount of
astigmatism that can be neutralized with a spherocylinder
correction within the 3-mm pupil zone). Asphericity is the
corneal asphericity inside the 4.5-mm diameter zone.
Corneal uniformity index is a measure of the distortion of
the corneal surface within the 3-mm pupil zone expressed
as a percentage. The predicted corneal acuity estimates the
predicted acuity within the 3-mm pupil zone, ranging from
20/10 to 20/200, assuming that the cornea is the limiting
factor in the visual system. Interpair topographic pattern
VOL. 127, NO. 2 CORNEAL TOPOGRAPHY AND
similarities were evaluated by scoring the same patterns
between twin pairs with a score of 1 and different patterns
with a score of 0.
Refractive data were analyzed by regression analysis of raw
data for each variable, as well as Student t test on intrapair
differences for each measurement. Topographic data were
entered into an Excel database and then converted for
statistical analysis by means of SAS statistical software. The
geometric means of the predicted corneal acuities were
calculated by the log of the inverse of the Snellen visual
acuities (logMAR visual acuity).9 Using analysis of variance
(general linear models analysis of variance), we compared the
interpair differences for the following values from the Holla-
day Diagnostic Summary: steep refractive power, flat refrac-
tive power, effective refractive power, total astigmatism
(vector differences), irregular astigmatism, asphericity, and
predicted corneal acuity (logMAR values). Fisher exact test
was used for corneal uniformity index. Chi-square test and
variance analysis were done to compare interpair pattern
similarity scores.
RESULTS
FORTY-TWO SETS OF TWINS WERE RECRUITED. TWENTY
monozygotic and 19 dizygotic sets of twins were enrolled in
the study. Ages for enrolled twins ranged from 12 to 73
years. Zygosity was readily determined in all but two sets of
twins by history, phenotypic characteristics, and survey
results. Two twin pairs were excluded due to undetermined
zygosity. One pair of monozygotic twins was excluded due
to previously diagnosed keratoconus. The spherical com-
ponent of refractive errors ranged from !1.25 to "9.50 D
(mean, "1.13 # 2.29 D) among monozygotic pairs and
!3.25 to "6.00 D (mean, "1.21 # 2.03 D) among
dizygotic pairs. (Data described are for the right eye only.
Left eye data were comparable for all variables.) Mean
intrapair differences of the spherical equivalent of refrac-
tion were less for monozygotic than for dizygotic twins
(0.38 # 0.065 vs 1.38 # 1.46; Student t test: P $ .001).
Regression analysis of the spherical equivalent showed a
significantly higher level of correlation for monozygotic
twins than for dizygotic twins (R 2 $ .957 vs .311) (Figure
1). The cylindrical component of refractive errors ranged
from 0 to 6.25 D (mean, 0.97 # 1.28 D) among monozy-
gotic pairs and 0 to 1.0 D (mean, 0.57 # 0.34 D) among
dizygotic pairs. Vector analysis as reported by Alpins10 was
performed. There was no significant mean intrapair differ-
ence between the monozygotic and dizygotic twins in
magnitude or direction of the astigmatic vector (Student t
test: P $ .46 and .36, respectively). Keratometric data
were also analyzed in the vertical and horizontal meridians.
There was no significant mean intrapair difference be-
tween the monozygotic and dizygotic twins in horizontal or
vertical meridians (Student t test: P $ .25 and .06,
respectively). Mean intrapair differences for axial lengths
REFRACTIVE VARIABLES IN TWINS 159
FIGURE 1. Regression analysis of spherical equivalent in monozygotic twins (left) and in dizygotic twins (right).

total astigmatism (vector analysis), and astigmatic axis. On

TABLE 1. Mean Intrapair Differences in Refractive the other hand, we found trends for lower differences in
Variables for Monozygotic and Dizygotic Twins dizygotic pairs for predicted corneal acuity and corneal
uniformity index.
Mean # SD

Monozygotic Dizygotic P
Refractive Variable Twins Twins Value*
DISCUSSION
Spherical component (D) 0.41 # 0.28 1.53 # 1.0 .001
Cylinder magnitude (D) 0.50 # 0.41 0.51 # 0.37 .46 THE AVERAGE LENGTH OF THE HUMAN EYE IS 18 MM AT
Cylinder meridian 39.1 # 3.24 42.4 # 28.46 .36 birth. The eye then undergoes two phases of axial growth:
(degrees)
a rapid infantile phase and a slower juvenile phase. This
Keratometric vertical (D) 0.52 # 0.55 0.92 # 0.88 .06
growth results in a myopic shift in refractive power of 16 to
Keratometric horizontal 0.70 # 0.69 0.90 # 0.98 .25
(D)
20 diopters. In a process referred to as emmetropization,
Axial length (mm) 0.39 # 0.29 0.76 # 0.58 .03 changes in lens power and corneal power compensate for
this myopic shift.11 A large study of emmetropic and
*Student t test. ametropic individuals showed a wide range of “normal
values” for each refractive variable. In most individuals,
these components correlate in such a way as to result in an
were less for monozygotic twins than dizygotic twins (axial emmetropic eye. Ametropia can develop in one of two
length: 0.39 vs 0.76 mm; Student t test: P $ .031) (Table ways. Component ametropia results when one component,
1). Regression analysis supported this higher correlation most often the axial length, falls outside the normal range.
(axial length: r 2 $ .838 vs .36) (Figure 2). Correlation ametropia results when each component falls
Corneal topographic data (power and meridian) in all within the normal range, but these components fail to
zones (3, 5, and 7 mm) also showed smaller mean differ- combine appropriately to focus an image on the retina. For
ences among monozygotic than dizygotic pairs, but the example, corneal curvature may be at the “steep end of the
differences were not statistically significant: Student t test normal spectrum,” while axial length is at the “long end of
in 3-mm and 7-mm zones (3-mm: magnitude [P $ .26], normal,” combining to form a myopic eye.1,12 Whether one
direction [P $ .18]; 7-mm: magnitude [P $ .07], or all of these components, or their relationships, are
direction [P $ .40]). However, the 5-mm zone showed genetically determined, or whether an individual’s envi-
significant mean intrapair difference between monozygotic ronment plays a significant role, has been a point of
and dizygotic twins in magnitude but not in direction speculation.
(magnitude, Student t test: P $ .006, direction, Student Current theories for the cause of myopia include the
t test: P $ .21) (Table 2). There were no statistically “use-abuse” theory and “biologic” or genetic theory. The
significant differences between the monozygotic and dizy- use-abuse theory stresses that environmental factors such
gotic twins in the intrapair comparison of topographic map as accommodation, convergence, diet, and education con-
patterns (Table 3). For the Holladay Diagnostic Summary tribute to myopia.13 The biologic theory, on the other
data, mean intrapair differences were significantly lower for hand, stresses that a genetic predisposition plays an impor-
monozygotic twins for irregular astigmatism and for dizy- tant role in determining myopia.3,14 –17 In our present
gotic twins for asphericity (Table 4). There were trends for study, the spherical equivalent of refraction shows higher
lower differences for monozygotic pairs for steep refractive correlation for monozygotic twins than for dizygotic twins.
power, flat refractive power, effective refractive power, There have been several other studies, including early

160 AMERICAN JOURNAL OF OPHTHALMOLOGY FEBRUARY 1999

FIGURE 2. Regression analysis of axial length in monozygotic twins (left) and in dizygotic twins (right).

TABLE 2. Mean Intrapair Differences in Effective Astigmatic Vectors from 3-, 5-, and 7-mm
Corneal Topographic Zones for Monozygotic and Dizygotic Twins

Mean # SD

Zone Monozygotic Dizygotic

(mm) Variable Twins Twins P Value*

3 Magnitude 1.41 # 1.11 1.59 # 0.84 .27

3 Meridian 45.9 # 25.5 37.9 # 28.5 .18
5 Magnitude 1.74 # 1.56 3.69 # 2.81 .006
5 Meridian 35.9 # 15.2 32.6 # 8.3 .21
7 Magnitude 1.49 # 0.96 1.02 # 0.95 .07
7 Meridian 43.9 # 28.2 41.4 # 31.0 .40

*Student t test.

TABLE 3. Classification of Corneal Topographic Maps in Monozygotic and

Dizygotic Twin Pairs

Monozygotic Twins Dizygotic Twins

(no. [%]) (no. [%])

Same Different Same Different

Map Type Pattern Pattern Pattern Pattern %2 P Value

Axial 10 (50) 10 (50) 11 (58) 8 (42) 0.244 .621

Local radius of curvature 12 (60) 8 (40) 9 (47) 10 (53) 0.626 .429
Refractive 12 (60) 8 (40) 14 (74) 5 (26) 0.821 .365
Profile difference 14 (70) 6 (30) 10 (53) 9 (47) 1.242 .265
Total, all patterns 48 (60) 32 (40) 44 (58) 32 (42) F $ 1.61; P $ .3231*

*General linear models analysis of variance.

studies by Sorsby and associates, the Finnish Twin Cohort
Studies, that had similar findings.3,16 –18 These results
suggest that there is a strong genetic determinant of the
spherical component of refractive error.
Our present study also compared the axial length differ-
ences between monozygotic and dizygotic twins. The
results show that there is a statistically significant intrapair
difference between the two groups, suggesting a genetic
contribution toward axial length. Similarly, high herita-
bility of axial length was reported by Sorsby and associates
in 1962.3 Unlike Sorsby and associates’ original study, in
which the axial length was calculated on the basis of
Gullstrand’s schematic eye, in our study axial length was
measured and therefore an independent variable. Several
experimental and clinical reports have demonstrated an
increase in axial length caused by visual deprivation from
cataract, vitreous hemorrhage, or corneal opacification,
especially in the early months of life.19 –21 However, when

VOL. 127, NO. 2 CORNEAL TOPOGRAPHY AND REFRACTIVE VARIABLES IN TWINS 161
 TABLE 4. Intrapair Differences in Variables of Holladay Diagnostic Summary for
Monozygotic and Dizygotic Twins

Mean # SD Variance Analysis

Monozygotic Twins Dizygotic Twins F* P

Steep RP (D) 1.05 # 0.97 1.11 # 0.96 1.15 .77

Flat RP (D) 0.79 # 0.83 1.05 # 0.97 1.37 .51
Effective RP (D) 0.87 # 0.88 1.04 # 0.85 1.05 .92
Astigmatism (D) 0.88 # 0.89 1.06 # 0.96 1.18 .73
Irregular astigmatism (D) 0.03 # 0.04 0.07 # 0.20 17.78 .000
Asphericity 0.37 # 0.50 0.19 # 0.19 6.92 .0001
Predicted corneal acuity 0.15 # 0.13 0.12 # 0.11 1.50 .398
Corneal uniformity index 6.50 # 11.36 4.73 # 6.96 — .354†

RP $ refractive power.
*General linear models analysis of variance.
†
Fisher exact test (two tailed).

there is no visual deprivation, there seems to be a strong
genetic predisposition toward axial length.
Unlike the refractive spherical equivalent, astigmatism
did not show a genetic predisposition in our study. Astig-
matism is a vector quantity that has both magnitude and
direction. Mean intrapair differences did not show a
statistically significant difference between monozygotic
and dizygotic twins in either magnitude or direction. This
suggests an environmental contribution toward the cylin-
drical component of refractive error. It is interesting to
note that individuals who have a propensity toward eye
rubbing, such as atopic patients, have a higher incidence of
keratoconus.22 However, most patients in the current study
had relatively low astigmatism, and the reliability and the
accuracy of manifest refraction in the assessment of mild
astigmatism is questionable. Unlike other twin studies, in
our study astigmatism was analyzed by vector analysis,
which takes into account both the magnitude and direc-
tion of the astigmatic vector. Our results concur with those
of the previous twin studies, including the Finnish Twin
Cohort Study and that of Lyle, which support an environ-
mental contribution to astigmatism.23,24
In addition, computerized videokeratography allowed
more accurate and complete information on corneal
power. We found few differences in the intrapair compar-
ison of the corneal topographic data for the monozygotic
and dizygotic pairs. The only intrapair values that were
lower for the monozygotic twins were corneal curvature at
the 5-mm zone and vector analysis of total astigmatism
from the refractive map. On the other hand, asphericity
differences were significantly lower for the dizygotic twins.
Possible explanations for our failure to find more differ-
ences between the monozygotic and dizygotic pairs include
(1) inadequate statistical power, (2) exclusion of contact
lens wearers from the study (they might have had higher
amounts of regular and irregular astigmatism that in turn
might have accentuated differences between pairs), or (3)
overriding environmental factors, which presumably ex-
plain the lower difference in asphericity for the dizygotic
pairs.
In conclusion, the results of this twin study suggest
genetic predisposition toward axial length and spherical
equivalent of refractive error. However, environmental
and other factors may play a significant role in determining
astigmatism. The use of monozygotic and dizygotic twins
provides a powerful tool for analyzing genetic and envi-
ronmental factors on a disease cause or a trait. However,
one should keep in perspective that there exists a greater
intrapair correlation for environmental factors for monozy-
gotic twins compared to dizygotic twins. This predisposes
toward a genetic bias, and our results might therefore
overestimate the genetic contribution to refractive error
and axial length, and underestimate the environmental
influence of these and other biometric variables.16
REFERENCES
1. Sorsby A. Biology of the eye as an optical system. In Duane
TD, Jaeger EA, editors. Clinical Ophthalmology. Philadel-
phia: JB Lippincott, 1989;34:1–17.
2. Raviola E, Wiesel TN. An animal model of myopia. N Engl
J Med 1985;312:1609 –1615.
3. Sorsby A, Sheridan M, Leary GA. Refraction and its com-
ponents in twins. Med Res Council Special Rep Ser 1962;
303:1– 43.
4. Sarna S, Kaprio J, Sistonen P, Koskenvuo M. Diagnosis of
twin zygosity by mailed questionnaire. Hum Hered 1978;28:
241–254.
5. Koch DD, Foulks GN, Moran CT, Wakil JS. The EyeSys
Corneal Analysis System: analysis of accuracy and reproduc-
ibility of first-generation prototype. Refract Corneal Surg
1989;5:424 – 489.
6. Budak K, Friedman NJ, Koch DD. Classification of corneal
topography in myopic eyes based on axial, instantaneous, and
refractive maps. J Cataract Refract Surg. Forthcoming.

162 AMERICAN JOURNAL OF OPHTHALMOLOGY FEBRUARY 1999

 7. Holladay J. Corneal topography using the Holladay Diagnos-
tic Summary. J Cataract Refract Surg 1997;23:209 –221.
8. Holladay JT, Cravy TV, Koch DD. Calculating the surgically
induced refractive change following ocular surgery. J Refract
Corneal Surg 1992;18:429 – 443.
9. Holladay JT, Prager TC. Mean visual acuity. Am J Ophthal-
mol 1991;111:372–374.
10. Alpins NA. A new method of analyzing vectors for changes
in astigmatism. J Cataract Refract Surg 1993;19:524 –533.
11. Sorsby A, Sheridan M, Leary GA. Refraction and its com-
ponents during growth of the eye after the age three. Med
Res Council Special Rep Ser 1961;301:1–18.
12. Sorsby A, Leary GA, Richards MJ. Correlation ametropia
and component ametropia. Vision Res 1962;2:309 –313.
13. Richler A, Bear JC. Refraction, near work and education.
Acta Ophthalmol 1980;58:468 – 478.
14. Karlsson JL. Concordance rates for myopia in twins. Clin
Genet 1974;6:142–146.
15. Karlsson JL. Evidence for recessive inheritance of myopia.
Clin Genet 1975;7:197–202.
16. Teikari J, O’Donnell JJ, Kaprio J, Koskenvuo M. Genetic and
environmental effects on oculometric traits. Optom Vision
Sci 1989;66:594 –599.
Authors Interactive"
We encourage questions and comments regarding this article via the Internet
on Authors Interactive" at http://www.ajo.com/ Questions, comments, and
author responses are posted.
VOL. 127, NO. 2 CORNEAL TOPOGRAPHY AND
17. Teikari JM, Kaprio J, Koskenvuo MK, Vannas A. Heritability
estimate for refractive errors: a population-based sample of
adult twins. Genet Epidemiol 1988;5:171–181.
18. Teikari JM, Koskenvuo M, Kaprio J, O’Donnell J. Study of
gene-environment effects on development of hyperopia: a
study of 191 adult twin pairs from the Finnish Twin Cohort
Study. Acta Genet Med Gemellol 1990;39:133–136.
19. Irving EL, Sivak JG, Callender MG. Refractive plasticity of
developing chick eye. Ophthalmic Physiol Opt 1992;12:
448 – 456.
20. Miller-Meeks M, Bennett SR, Keech RV, Blodi CF. Myopia
induced by vitreous hemorrhage. Am J Ophthalmol 1990;
109:199 –203.
21. Rasooly R, BenEzra D. Congenital and traumatic cataract:
the effect on axial length. Arch Ophthalmol 1988;106:
1066 –1068.
22. Krachmer JH, Feder RS, Belin MW. Keratoconus and related
noninflammatory corneal thinning disorders. Surv Ophthal-
mol 1984;28:293–322.
23. Lyle WM. The inheritance of corneal astigmatism [thesis].
Bloomington, Indiana, Indiana University, 1965:47–50.
24. Teikari JM, O’Donnell JJ. Astigmatism in 72 twin pairs.
Cornea 1989;8:263–266.
REFRACTIVE VARIABLES IN TWINS 163