Heterotopic Muscle Pulleys or Oblique

Muscle Dysfunction?
Robert A. Clark, MD, -aJoel M. Miller, PhD, -cArthur L. Rosenbaum, MD? and Joseph L. Demer, MD, P h I ~

Introduction: The description of connective tissue sleeves that function as pulleys for the rectus extraocular
muscles (EOMs) suggests that abnormalities of EOM pulley position might provide a mechanical basis for some
forms of incomitant strabismus. Pulleys determine the paths and thus the pulling directions of EOMs. Methods:High-
resolution magnetic resonance images spanning the orbits were obtained in primary position, upgaze, and
downgaze for each subject. Paths of the EOMs were measured with reference to the orbital center and permitted
inference of pulley locations. Results:Data from 18 orbits of orthotropic subjects defined means and SDs of normal
EOM pulley coordinates. Eight patients, aged 17 to 60 years, had heterotopic EOM pulleys, defined as displaced at
least 2 SDs from normal. We found one to eight heterotopic pulleys (considering both orbits)in each of four patients
who had been diagnosed with marked superior oblique (SO) overaction and mild to marked inferior oblique (10)
underaction. Each patient had superior mislocation of at least one lateral rectus pulley by 1.8 to 4.9 ram. Three
patients diagnosed with mild to moderate I0 overaction and mild to moderate SO underaction in only one orbit had
one to three heterotopic EOM pulleys. Each of those patients had at least one lateral rectus pulley inferiorly
dislocated by 1.9 to 4.9 ram. The final patient, who was diagnosed with mild I0 underaction and normal SO function
bilaterally, had bilateral superior mislocation of the medial rectus pulleys by greater than 2 mm. Computer
simulations using the Orbit program (Eidactics, San Francisco)incorporating individually measured pulley positions
reproduced the clinical patterns of incomitant strabismus in all cases without postulating abnormalities of oblique
muscle innervation or contractility. Conclusion: Heterotopic EOM pulleys can cause patterns of incomitant
strabismus that have been attributed to oblique muscle dysfunction. Even isolated mislocations of less than 2 ram,
coupled with smaller mislocations of the other pulleys, can produce the clinical appearance of bilateral oblique
dysfunction. Pulley heterotopy should be considered in the differential diagnosis of incomitant strabismus and
oblique dysfunction. (J AAPOS 1998;2:17-25)

he combination of modern orbital histologic tissue sleeves stabilize EOM bellies relative to the orbit,

T studies and high-resolution magnetic resonance

imaging (MRI) in alert humans has expanded our
knowledge of the mechanics of the rectus extraocular
muscles (EOMs) and their associated connective tissues.
Serial cross-sectional histologic studies have shown that
each rectus EOM passes through a connective tissue sleeve
composed of collagen, elastin, and smooth muscle in the
region of posterior Tenon's capsule, just behind the
equator of the globe. 1-3MRI has demonstrated that these
From the Departments of Ophthalmology ~and Neurology, ~ University of California, Los
Angeles, and Smith-Kettlewell Eye Research Institute, c San Francisco. R. A. C. is a
Rosalind W. Alcott Fellow.
Supported by grants from the National Eye Imtitnte: consortium grant No. EY-08313
(y. L. D., J. M. M.) and coregrants No. EY-00331 (Department of Ophthalmology, Uni-
versity of California, Los Angeles) and EY-068 8 3 (Smith-Kettlewell Eye ResearchInstitute).
Presented at the annual meeting of the American Associationfor Pediatric Ophthalmology
and Strabismus, Charleston, South Carolina, April 1997, under the title "HeterotopicRectus
Extraoadar Muscle Pulleys Simulate Oblique Mzacle Dysfunction."
Reprint requests: Joseph L. Demer, MD, PhD, Jules Stein Eye Institute, 100 Stein
Plaza, University of California, Los Angeles, Los Angeles, CA 90095-7002.
Copyright © 1998 by the A~rican Associationfor Pediatric Ophthalmology and Strabi~ma.
I091-8531/98 5Y.00 + 0 75/1/85163
permitting only the insertional ends of the muscles to move
with globe rotation. 4,5Effects of this suspension are clearly
seen in orbital images after muscle transposition surgery in
which large transpositions of the anterior tendon insertion
fail to shift the posterior EOM path. 6,7 The tissue sleeves
function as mechanical pulleys, becoming the effective
origins of EOM action.
We have measured the normal positions of all the EOM
pulleys and found them to be stereotypic in normal
subjects, as would be expected for important mechanical
structures.i, 4, s In addition, significantly abnormal pulley
positions have been shown to cause incomitant strabis-
musfi, 8Abnormal radial displacements of pulleys (as might
be associated with a large or small orbit) do not appear to
affect binocular alignment, but abnormalities of pulley
location perpendicular to a muscle's plane of action (i.e.,
vertical displacement of horizontal rectus muscle pulleys or
horizontal displacement of vertical rectus muscle pulleys)
can cause incomitant strabismus. 5 These findings suggest
that heterotopic pulleys may be an important cause of
incomitant strabismus. 5

Journal ofAAPOS February 1 9 9 8 17


18 Clark et al.
FIG. 1. Coronal plane MRI of normal right orbit demonstrates the
angular correction used to rotate all orbits to bring interhemispheric
fissure of brain to vertical position (as defined by magnetic field
coordinates of MRI scanner). SR, Superior rectus muscle; LR, lateral
rectus muscle; IR, inferior rectus muscle; MR, medial rectus muscle;
SO, superior oblique muscle.
Determining which subjects with incomitant strabismus
may have heterotopic pulley positions currently requires
high-resolution imaging of the orbit and analysis of
posterior EOM paths. This study was designed to
determine which clinical parameters may be important in
predicting heterotopic pulleys. We analyzed cases that
would ordinarily be diagnosed as oblique muscle dysfunc-
tions and interpret them as something quite different. We
will therefore describe alignment "patterns" in terms that
are neutral with respect to etiology.
SUBJECTS AND METHODS
Ten normal volunteers were recruited by advertisement
and were examined to verify normal ocular motility and the
absence of strabismus. Each subject was given a Hess
screen test to quantify binocular alignment in 21 fixation
positions over a +30-degree field for each eye. After
informed written consent was obtained according to a
protocol conforming to the Declaration of Helsinki and
approved by the Human Subject Protection Committee at
the University of California, Los Angeles, these subjects
underwent high-resolution MRI scanning with use of a
superconducting 1.5 T General Electric Signa (Milwau-
kee, Wisc.) or Picker Vista (Cleveland, Ohio) scanner by
techniques described in detail elsewhere.4-6 Briefly, sub-
jects' heads were stabilized with the axis of the scanned
orbit vertical with use of foam cushions and tape. Then a
surface coil was placed over the scanned orbit, and multiple
Journal ofAAPOS
Volume 2 Number 1 February 1998
contiguous images 3 mm in thickness were obtained with a
256 x 256 matrix over an 8 or 10 cm square field of view,
giving pixel resolutions of 312 or 390 gm, respectively.
Fixation targets were provided inside the scanner magnet.
Images of primary gaze, upgaze, and downgaze were
obtained for all subjects.
The digital MRI images were transferred to Macintosh
computers (Apple Computer, Cupertino, Calif.), con-
verted into eight-bit tagged image file format with use of
locally developed software and quantified with use of the
program National Institutes of Health (NIH) Image
(Rasband W, NIH, available by file transfer protocol from
zippy.nimh.nib.gov or on floppy disk from the National
Technical Information Service, 5285 Port Royal Road,
Springfield, VA 22161, part No. PB95-500195GEI).
Images of left orbits were digitally reflected to the
orientation of right orbits to allow uniform analysis of
EOM positions.
Normalization of image position and orientation
facilitated quantitative comparisons across subjects. For all
orbit and muscle locations the area centroid, defined by
digitally tracing the outline of each structure and computed
by the area centroid algorithm in NIH Image, was used to
determine the location of each structure to subpixel
resolution. To normalize position in the scan plane, all
rectus EOM positions were translated to place the
coordinate origin at the area centroid of the orbit. The
magnetic field coordinates of the MRI scanner served as
references for rotational corrections. Rotation in the
coronal plane was normalized by rotating the image to
align the interhemispheric fissure of the brain with the
scanner-defined vertical meridian (Figure 1). Finally, to
normalize the anterior-posterior position, the image plane
3 mm anterior to the globe-optic nerve junction was
selected for analysis in each subject. A more anterior image
plane, nearer to the globe equator, would have better
intersected the densest pulley regions, but the flatness of
muscle tendons and the density of connective tissue make it
difficult to distinguish contours by which muscle or pulley
position in the slice plane could be judged. The "3 mm
anterior" image plane was the most anterior in which all
rectus EOM bellies could clearly be identified in every
subject (Figure 2). The average locations and SDs of the
area centroids of EOM bellies in this plane for normal
subjects were then calculated. Pulley positions were
estimated as the positions of EOM area centroids.
With use of the same image alignment and analysis, eight
strabismic subjects from a continuing clinical series were then
identified as having one or more EOM positions more than 2
SDs from normal. Previous work has demonstrated that gaze
does not significantly affectpulley position, allowing accurate
determination of pulley position in eyes that are deviated
from primary position either from eccentric gaze or
strabismus? Only pulley displacements perpendicular to
each EOM's plane of action were considered,s In these
strabismic subjects we recorded age; duration of symptoms;
Journal ofAAPOS
Volume 2 Number 1 Febrnary 1998
FI6.2. Coronal MRI scans of right orbit of normal subject in primary position. Image plane 3 mm anterior
(left) to globe-optic nerve junction (right) was used for analysis. SR, Superior rectus muscle; LR,
lateral rectus muscle; IR, inferior rectus muscle; MR, medial rectus muscle; SO,superior oblique muscle;
ON, optic nerve; LPS, levator palpebrae superioris.
refractive error; best corrected visual acuity; clinical binocu-
lar alignment deviation in upgaze, primary gaze, and
downgaze at distance and in primary gaze at near; versions;
double Maddox rod measurement of torsion; and conven-
tional clinical diagnosis.
Pulley locations were inferred and binocular alignment
simulated with use of the Orbit 1.6 Gaze Mechanics
Simulation program 9 (Eidactics, San Francisco) running
on Macintosh computers. Orbit simulates binocular
alignment with use of static force balance equilibrium
equations that are based on orbital parameters such as
innervations, globe dimensions, and EOM insertions,
lengths, stiffness, pulley positions, and contractile forces.
The program then calculates the behavior of the EOMs
and globes on the basis of equations and methods given, in
part, by Robinson, TMMiller and Robinson, 11and Miller and
DemerJ 2 Pulley positions (based on histologic studies2) in
Orbit's description of a normal eye were taken as the
starting point. 5 Then the lateral-medial and superior-
inferior coordinates of Orbit's pulleys were altered to
match Orbit's simulated muscle paths to the paths observed
in the MRI scans. Small rectus muscle length changes were
then applied to account for horizontal deviations in
primary gaze (i.e., shortening the medial rectus muscles to
simulate esotropia). The investigator performing the
simulations did not know the observed pattern of deviation
for each subject.
RESULTS
Eighteen orbits in 10 normal subjects were analyzed,
providing 72 rectus EOMs. The results for the 3 mm anterior
image plane are summarized in Table 1. Two normal subjects
each had one EOM pulley displaced more than 2 SDs from
normal. This number is not unexpected on the basis of a
normal distribution of pulley coordinates. In one subject the
Clark et al. 19
TABLE 1. Position of normal rectus extraocular muscles
(millimeters from orbital center)
X ¥
Medial rectus 12.0 +0.6 0.6 +0.9
Superior rectus -1.4 +0.8 12.6 +0.7
Lateral rectus -11.9 +0.5 -0.6 +0.9
Inferior rectus 2.1 +0.9 -11.8_+0.9
Primary position of rectus EOMs in coronal image plane 3 mm anterior to globe-
optic nervejunction, expressed as coordinates relative to orbital center. The x
coordinate represents horizontal position, with positive values representing
medial displacement; ycoordinate represents vertical position, with positive
values representing superior displacement. The +errors shown represent 1 SD.
right medial rectus pulley was shifted inferiorly by 2.3 mm
from normal. In the other subject the right lateral rectus
pulley was shifted superiorly by 1.8 mm from normal. In both
subjects no other EOM pulley was displaced more than 1 SD
from normal. Neither subject had any clinical abnormality on
examination or Hess screen test. In computer simulation
using the measured pulley positions, the maximum computed
deviation in primary gaze for both subjects was 0.6 degrees
(1.1 PD) of horizontal deviation, 1.0 degree (1.8 PD) ofhyper
deviation, and 1.3 degrees of cyclotorsion, all well within the
range of binocular fusion, r3 The maximum difference in
horizontal deviation from upgaze to downgaze was 4.9
degrees (8.6 PD).
Sixteen orbits in eight strabismic subjects were analyzed.
Two subjects, No. 1 and No. 4, have been described
elsewhere.S,s The eight strabismic subjects could be dMded
into three groups on the basis of clinical information and
heterotopic pulley location; The first group of four subjects
each had "A" pattern strabismus (eyes deviated outward more
in downgaze than upgaze) and, on version examination,
 Journal of AAPOS
20 Clark et aL Volume 2 Number I February 1998

FIG. 3. Clinical photographs of versions of subject 4 in nine diagnostic
positions of gaze, demonstrating bilateral overdepression in adduction
("superior oblique muscle overaction") and underelevation in
adduction ("inferior oblique muscle underaction").
15,0 -
FIG. 5. Clinical photographs of versions of subject 6 in nine diagnostic
positions of gaze, demonstrating overelevation of left eye in adduction
("inferior oblique muscle overaction") and underdepression of left
eye in adduction ("superior oblique muscle underaction").
15.0

•r •
i! I
10.0 - 10.0 -
O SR
SR

E
E E
5.0- E 5.0-
$

o--~ [] 0 o

00.O-

LR MR 8 gE -I
.B 00
-5,0- ~ -5.0-
a a
>.
IR
-10.0 - IR
-I0.0-

• •
nmT
•'~0"~1
T
°*41
-15.0 i i -15.0 I
-15.0 -10.0 -1.o 0'.0 ;.0 10.0 15.0 -15.0 100 -;0 010 ;0 '
10.0 15.0
X Deviation from the Orbital Center, mm X Deviation from the Orbital Center, mm

• Normal RectusPulley Positions • Subject 30D • Normal Rectus Pulley Positions • Subject 6 0 S

0 Subject 10D • Subject 3 0 S " 0 Subject 50D • Subject 70D

• Subject 10S =~ Subject 40D • Subject50S • Subject7 0 S

[] Subject 20D • Subject 4 0 S [] Subject 60D

• Subject 2 0 S
RG. 4. Average positions (relative to right orbital center and viewed
as if facing subject) of centroids of rectus EOMs for primary gaze in
coronal image plane 3 mm anterior to globe-optic nerve junction for
normal and for strabismic subjects with overdepression in adduction
("superior oblique muscle overaction"). Left orbits have been digitally
transposed to configuration of right orbits to facilitate data analysis.
Error bands shown represent +_2 SDs (95% confidence intervals)
around normal means. SR, Superior rectus muscle; LR, lateral rectus
muscle; MR, medial rectus muscle; IR, inferior rectus muscle.
FIG. 6. Average positions (relative to right orbital center and viewed as
if facing subject) of centroids of rectus EOMsfor primary gaze in coronal
image plane 3 mm anterior to globe-optic nerve junction for normal
subjects and for strabismic subjects with unilateral underdepression in
adduction ("SO underaction'). Left orbits have been digitallytransposed
to configuration of right orbits to facilitate data analysis. Error bands
shown represent + 2 SDs (95% confidence intervals) around normal
means. SR,Superior rectus muscle; LR,lateral rectus muscle; MR, medial
rectus muscle; IR, inferior rectus muscle.

showed marked overdepression in adduction ("superior
oblique overaction") with marked underelevation in adduc-
tion ("inferior oblique underaction") (Figure 3, Table 2).
These subjects also all had at least one lateral rectus muscle
pulley displaced superiorly more than 2 SDs from normal
(Table 3), from 1.8 mm to 4.8 mm above normal (Figure 4).
Three of the four subjects had several other
heterotopic pulleys and had congenital strabismus.
Subject 4 was seen late in life with what was diagnosed as
bilateral inferior oblique palsy. For subject 4, only the
right lateral rectus muscle pulley was displaced superi-
orly more than 2 SDs from normal, but the other EOM
pulleys in the right orbit were all displaced more than 1
SD from normal. Computer simulation of the measured
pulley positions, without any changes in oblique muscle
function, produced remarkably close agreement in the
Journal of M P O S
Volume 2 Number I February 1998 Clark et al. 21

TABLE 2. Clinical characteristics of strabismic subjects

Deviation in Deviation in Deviation in Deviation at Version Double Clinical
Subject Age(yr) Duration up-gaze primary down-gaze near abnormalities Maddox red diagnosis

1 30 Congenital X = 4, XT = 4, XT = 35, RH = 4 SO +4 OU, Not done "A" pattern XT with

RH = 4 RHT = 4 RHT = 15 LID -4 SO overaction OU
2 36 Congenital 2 XT Flick XT, 10 XT, 30 XT SO +40U, <5° intorsion "A" pattern XT with
5 LHT 8 LHT LR +1 OU, RIO -1 SO overaction OU
3 23 Congenital 3 LXT, 2 LXT, 18 LXT, 3 XT, RSO +2, 12° intorsion "A" pattern XT with
3 LHT 3 LHT 4 LHT 3 LHT RIO -2, LLR -1/2 SO overaction OD
4 50 3 yr 35 ET 30 ET, 16 ET 30 ET SO +30U, 1-3° intorsion Bilateral IO palsy
6 RHT I0 -30U
5 60 16 months 18 X(T), 6 X(T), 25 LHT Not done LIO +1, 5-6° extorsion LSO palsy
14 LH(T) 16 LH(T) LSO -2
6 36 7 months 8 RHT 16 RHT 12 RHT Ortho RSO -~/2, 10° extorsion RSO palsy
RIO+I
7 46 12-15 yr 10 LHT 10 LHT Ortho 6-8 LHT RSR -2, 5-10° extorsion RSR palsy
LIO +2, LSO -2
8 17 2-3 yr 10 E(T), 20 E(T), 30 E(T), 25 ET, LR -1/20U, Not done "V" pattern ET without
4 RH(T) 4 RH(T) 4 RH(T) 4 RHT IO -1 OU IO overaction
Summaryof the clinical information of all strabismic subjects. Deviationsare measuredin prism diopters. Versionsare graded on scale of -4 to +4, with a grade of 0 as
normal, -4 as maximalunderaction,and +4 as maximaloveraction. Onlyabnormal versionsare listed; X, exophoria;RH, right hyperphoria;XT, exotropia; RHT,right
hypertropia; SO, superioroblique muscle; OU, both eyes; LIO, left inferior oblique muscle; LHT, left hypertropia;LR, lateral rectus muscle; RIO, right inferior oblique muscle;
LXT, left exotropia;RSO, right superior oblique muscle; LLR, left lateral rectus muscle; ET, esotrepia;/0, inferior oblique muscle;X(T), intermittent exotropia;LH(T),
intermittent left hypertropia; LIO, left inferior oblique muscle; LSO, left superioroblique muscle; RSR, right superior rectus muscle; E(T)intermittent esotropia; RH(T),
intermittent right hypertropia.

TABLE 3. Pulley position deviations in strabismic subjects

Subject MR deviation (OD/OS) SR deviation (OD/OS) LR deviation (OD/OS) IR deviation (OD/OS)
1 Down 3.2*/down 4.8* Medial 2.0*/medial 2.6* Up 3.5*/up 4.8* Lateral 2.7*/lateral 5.0*
2 Down 2.7*/up 0.7 Medial 1.3/lateral 0.9 Up 3.1 */up 0.7 Lateral 3.1*/lateral 1.2
3 Down 1.1/up 0.8 Medial 1.2/medial 0.4 Up 2.4*/up 1.9" Lateral 3.6*/lateral 1.7
4 Down 1.1/down 0.4 Medial 1.2/medial 0.9 Up 1.8*/down 1.0 Lateral 1.4/lateral 0.5
5 Down 2A*/up 0.1 Lateral 1.4/lateral 1.8 Down 4.0*/down 4.9* Medial 0.3/medial 0.7
6 Up 0.5/up 0.9 Lateral 2.1*/lateral 2.8* Down 1.9*/down 0.6 Medial 2.0*/medial 2.3*
7 Down 0.1/up 1.3 Medial 1.0/lateral 0.2 Down 1.2/down 3.5* Medial 1.1/medial 1.2
8 Up 2.5*/up 2.1" Medial 1.3/no change Up 0.3/up 0.6 Medial 1.1/lateral 0.2
Values are millimeters of deviationfrom normal. Summaryof pulley position deviationsfrom normal in all strabismicsubjects. Onlydeviationsperpendicularto each
muscle's field of action are listed.
*Deviations more than 2 SDs from normal.

amount of "A" pattern deviation seen in each subject
(Table 4), with the exception of subject 1, whose
simulated deviation substantially overestimated the
measured deviation. Before analysis, however, subject 1
underwent bilateral superior oblique tenotomies, which
may have reduced the original amount of the "A"
pattern deviation.
Each of the second group of three Subjects had
incomitant hypertropia and, on version examination,
showed moderate underdepression in adduction ("supe-
rior oblique muscle underaction") with moderate
overelevation in adduction ("inferior oblique muscle
overaction") in one orbit (Figure 5, Table 2). One subject,
subject 5, also had a superimposed clinically significant
"V" pattern exotropia (eyes deviated inward more in
upgaze than downgaze), whereas the other two subjects
demonstrated a small "V" pattern on Hess screen test.
These subjects all had at least one lateral rectus muscle
pulley displaced inferiorly more than 2 SDs from normal
(see Table 3), from 1.9 to 4.9 m m below normal (Figure 6).
All three subjects had onset of symptoms in adulthood.
Computer simulation of the measured pulley positions,
 Journal ofAAPOS
22 Clark et al. Volume 2 Number 1 February 1998

20.0

15.0 - •O
T
I-,-•..-I
,..L SR
10.0 -
E
E
*E
5.0-
0
Fill. 7. Clinical photographs of versions of subject 8 in nine diagnostic
positions of gaze, demonstrating bilateral underelevation in
adduction ("inferior oblique muscle underaction") with normal
O
0.0- & I'-,'-I
1 MR
depression in adduction. E /
LR
-5.0 -

a
without any changes in oblique muscle function, produced >-
IR
-10.0 -
remarkably close agreement in the amount of"V" pattern T
deviation seen in each subject (see Table 4), with the
exception of subject 6, whose simulated deviation overes- -15.0-

timated his actual deviation.

The final subject, subject 8, had an unusual "V" pattern -20,0 i , ; , , , I
esotropia with underelevation in adduction ("inferior -20.0 -15.0 -10.0 - .0 0.0 5 0 10.0 15.0 20.0

oblique muscle underaction") (Figure 7, Table 2). Both X Deviation from the Orbital Center, mm

medial rectus muscle pulleys were displaced superiorly • Normal Rectus Pulley Positions

more than 2 SDs from normal (see Table 3), from 2.1 to
2.4 mm above normal (Figure 8). Computer simulation of
the measured pulley positions, without any changes in
oblique muscle function, underestimated the amount of
"V" pattern deviation clinically seen (see Table 4). After
the subject underwent uncomplicated bilateral medial
rectus muscle recession of 4.5 mm with infraplacement 4.0
mm, the "V" pattern collapsed and the subject was
orthotropic on cover-uncover testing. He complained of
severe tilting and diplopia, however, and on double
Maddox rod testing was found to have 15 degrees of
relative extorsion postoperatively. Computer simulation
of the actual surgery with the measured pulley positions
predicted the generation of greater than 10 degrees of
extorsion postoperatively.
In both normal and strabismic subjects the cross-
sectional area of the superior oblique muscle was measured
in primary gaze, upgaze, and downgaze with use of analysis
techniques described elsewhere. 14-16No strabismic subject
had a superior oblique muscle palsy by MRI criteria, with a
range of superior oblique muscle maximum cross-sectional
area in primary gaze of 0.16 cm z to 0.33 cmz (normal
average 0.19 cm2) Is and minimum contractile change in
cross-sectional area from 23 degrees upgaze to 23 degrees
downgaze of 0.03 cm2.
DISCUSSION
Many clinical cases show that large rectus EOM pulley
abnormalities can cause incomitant strabismus. Cranio-
synostosis syndromes (i.e., Apert, Crouzon, and Pfeiffer
syndromes) are associated with small, laterally rotated
orbits, abnormally located EOMs, and a marked "V"
pattern exotropia. 17 Similarly, "heavy eye syndrome" is
associated with high axial myopia and a large inferior
O Subject8 0 D
• Subject8 0 S
FI6.8. Average positions (relative to right orbital center and viewed
as if facing subject) of centroids of rectus EOMs for primary gaze in
coronal image plane 3 mm anterior to globe-optic nerve junction for
normal subjects and for subject 8. Left orbits have been digitally
transposed to configuration of right orbits to facilitate data analysis.
Error bands shown represent + 2 SDs (95% confidence intervals)
around normal means. SR,Superior rectus muscle; LR, lateral rectus
muscle; MR, medial rectus muscle; IR, inferior rectus muscle.
displacement of the lateral rectus muscle, resulting in
esotropia and hypotropia as the normal abducting action
of the lateral rectus muscle is converted to depression. TM
This study demonstrates that small mislocations (<2 nun)
of the rectus EOM pulleys can also cause incomitant
strabismus. Computer modeling indicates that these pulley
displacements alone can account for the incomitancies,
without supposing any abnormalities of the oblique muscles.
Although our sample is small, three distinct patterns of
incomitant strabismus were associated with heterotopic
pulley position.
The first pattern consists of overdepression and
underelevation in adduction and has been called "marked
superior oblique muscle overaction with inferior oblique
muscle underaction." The pathophysiologic mechanisms
of this common syndrome may be complex. For cases in
which muscle overaction is really involved, causes might
include abnormal innervation resulting in increased muscle
stimulation or an abnormally powerful muscle that
responds excessively to normal innervation. A short, stiff
("contractured") muscle can produce somewhat similar
Journal ofAAPOS
Volume 2 Number 1 February 1998
effects. Clinical examples exist, however, including subject
1 in this study, in which complete superior oblique muscle
tenotomyis ineffective in relieving the presumed overaction.
Such cases may be better explained by noting that
abnormal superior displacement of a lateral rectus muscle
pulley can generate the clinical pattern called "superior
oblique muscle overaction." Computer simulation shows
that a superiorly displaced lateral rectus muscle pulley can
generate "A" pattern strabismus. Two lines of clinical
reasoning also support a potential role for a superiorly
displaced lateral rectus muscle pulley in this syndrome. First,
superior transposition of the lateral rectus muscle tendon can
correct "V" pattern strabismus. Therefore superior transpo-
sition of the lateral rectus muscle tendon in a normal subject
would be expected to give the lateral rectus muscle excessive
passive tension in downgaze, generating "A" pattern
strabismus. Second, during version examination a superiorly
displaced lateral rectus muscle pulley would elevate the globe
during abduction to the affected side. To maintain fixation by
the affected eye, the inferior rectus muscle on the affected
side would need to contract to maintain constant vertical eye
position. Through Herring's law, the inferior rectus muscle
on the other, adducting eye would also receive this
innervation and contract. The nonfixating, adducting eye
would then exhibit excessive depression, the clinical hallmark
of "superior oblique muscle overaction."
A similar analysis can be made regarding the second
pattern of incomitant strabismus, called "superior oblique
muscle underaction with inferior oblique muscle
overaction." Is superior oblique muscle palsy the cause of
this incomitant strabismus? Recent studies with high-
resolution MRI demonstrate that one third to one half of
presumed superior oblique muscle palsies do not have the
expected atrophy and impaired contractility of the involved
superior oblique muscle belly) 9, 20 Computer simulation
shows that an inferiorly displaced lateral rectus muscle
pulley can generate "V" pattern strabismus and apparent
superior oblique muscle underaction without any contrac-
tile weakness of the superior oblique muscle. Likewise,
during version examination, an inferiorly displaced lateral
rectus muscle pulley would depress the globe during
abduction to the affected side. T o maintain fixation, the
superior rectus muscle on the affected, fixating side would
need to contract to maintain constant vertical eye position.
Through Herring's law, the superior rectus muscle of the
opposite, nonfixating adducting eye would also contract.
The adducting eye would then exhibit excessive elevation,
the clinical hallmark of "inferior oblique muscle overaction."
An argument might be made that the change in position
of the pulleys could be a secondary phenomenon related to
torsion of the globes. 2~ For example, during an acute
superior oblique muscle palsy, the resulting extorsion of
the globe might be supposed to drag the lateral rectus
muscle pulley inferiorly, where it might hypothetically
remain after the acute palsy had resolved. If simple
extorsion were the explanation, however, all the EOM
Clark et aL 23
TABLE 4. Variation in horizontal alignment in strabismic subjects from
up-gaze to downgaze (positive values more exotropic on down-gaze)
Actual change in Simulated change in
horizontal deviation horizontal deviation
Subject (degrees) (degrees)
1 11 (19.4 PD) 30 (57.7 PD)
2 6 (10.5 PD) 7 (12.3 PD)
3 9 (15,8 PD) 9 (15.8 PD)
4 5 (10.5 PD) 6 (10.5 PD)
5 -5 (-8.8 PD) -7 (-12.3 PD)
6 -5 (-8.8 PD) -11 (-19.4 PD)
7 -7 (-12.3 PD) -9 (-15.8 PD)
8 -14(-24.9 PD) -7 (-12.3 PD)
Summary of observed versus simulated change in horizontal deviation from 30
degrees upgaze to 30 degrees downgaze in all strabismic subjects, as measured
by Hess screen. In subjects 1 through 4, horizontal deviation widened in
downgaze (i.e., increasing exotropia or decreasing esotropia), consistent with "A"
pattern strabismus. In subjects 5 through 8, horizontal deviation narrowed in
downgaze (i.e., decreasing exotropia or increasing esotropia), consistent with "V"
pattern strabismus.
pulleys should be displaced to the same degree, not just the
lateral rectus muscle pulley. In addition, recent experi-
mental data suggest that only the medial rectus muscle
pulley position is significantly affected by superior oblique
muscle palsy, and the magnitude of the effect is much less
than 2 SDs from the normal positionfl 2 Finally, from
simple geometric analysis an extorsion of 10 degrees in a
24 mm globe would generate less than 2 mm of
displacement of the anterior, tendinous insertion of the
EOMs. This is the maximum displacement of posterior
EOM bellies, supposing the pulleys have no intrinsic
stiffness and, so, displace freely with the EOM tendon.
However, even with muscle transposition surgery of more
than 6 mm, posterior muscle bellies do not displace
significantly from normal. 6 It is therefore unlikely that
torsion, which displaces muscle insertions less, could
significantly displace EOM pulleys.
An argument could also be made that true oblique
muscle dysfunction, not heterotopic pulleys, must be the
primary cause of most cases of clinical overelevation and
overdepression in adduction because surgery on oblique
muscles is effective in correcting the patterns of strabismus.
The fact that a surgery is successful, however, does not
imply that the surgery directly reversed the pathophysi-
ologic features that created the strabismus. Few would
argue, for example, that anterior placement of the medial
rectus tendon causes esotropia. Many successful surgeries
to correct esotropia, however, involve posterior reinsertion
of the medial rectus tendon. Similarly, recession of the
inferior oblique muscle in patients with overelevation in
adduction may correct the clinical problem by weakening a
muscle that elevates the eye in adduction, regardless of
whether the inferior oblique muscle itself is at fault.

24 Clark et al.
Additional arguments have been made that aberrant
innervation of the sixth nerve can cause "V" and "Y"
pattern strabismus, 23,24 using as evidence abnormal elec-
tromyographic recruitment of the lateral rectus muscle
during elevation in adduction. This explanation supposes
coinnervation of lateral rectus and superior rectus muscles.
We have no reason to think that this is not possible. We
offer another mechanism that could explain lateral rectus
muscle contraction during elevation, on the basis 0flateral
misplacement of the superior rectus muscle pulley, seen in
two of the three subjects with "V" pattern strabismus and
overelevation in adduction. If the superior rectus muscle is
laterally misplaced on the fixating eye, the medial rectus
muscle would need to contract to maintain alignment
during elevation. Through Herring's law the lateral rectus
muscle of the opposite, nonfixating, adducting eye would
also contract, giving rise to the abnormal recruitment of
muscle fibers during elevation.
The third pattern of incomitant strabismus, exhib-
ited by subject 8, is an uncommon "V" pattern esotropia
with underelevation in adduction. Typical "V" pattern
esotropia would be expected to be associated with
overelevation in adduction, resulting in the characteris-
tic excessive abduction in elevation. Subject 8 was
unusual also in that the medial rectus pulley, not the
lateral rectus pulley, was abnormally elevated. Given
the magnitude of the clinical deviation versus the much
smaller deviation predicted by computer simulation, it
is possible that other, unknown factors beside hetero-
topic pulley position may be contributing to his
incomitant strabismus. This highlights the fact that
properties of the pulleys, particularly the inferior
oblique pulley and its attachment to the inferior rectus
muscle 2s through Lockwood's ligament, 26require better
quantitative characterization in the Orbit model.
More interesting, however, is subject 8's highly
anomalous result after uncomplicated medial rectus
transposition surgery. On postoperative day 1 the patient
appeared to have a new-onset bilateral superior oblique
muscle palsy with 15 degrees of extorsion. Although the
exact insertions of the medial rectus tendons were not
measured intraoperatively, they certainly did not appear
clinically to be elevated 2 to 3 ram. Assuming the medial
rectus tendons inserted in the normal position, approxi-
mately 2.5 mm below the heterotopic medial rectus
pulleys, inferior transposition of the medial rectus tendons
by 4.5 mm effectively moved the insertions 7 mm below
the pulleys. Thus the medial rectus muscle was surgically
turned into an excyclotorter in both eyes, capable of
generating the high degrees of extorsion seen both
clinically and predicted by computer simulation.
Another clinical parameter of interest was age at onset
of symptoms. Subjects with congenital incomitant strabis-
mus had multiple markedly heterotopic pulleys. Other
subjects with adult-onset strabismus had fewer muscles
involved and smaller changes in pulley position. The
Journal ofAAPOS
Volume 2 Number 1 February 1998
resulting binocular deviations probably remain well
controlled through motor fusional mechanisms until an
event later in life disrupted binocular fusion and allowed
the underlying incomitance to become unmasked acutely.
Such events could include the formation of a cataract with
resultant decreased vision or anisometropia or an acute
illness that transiently impaired fusion. Then, because of
the imbalance inherent from the heterotopic pulley
positions, fusion could not be restored and incomitant
strabismus persisted.
Currently, it is unknown whether heterotopic pulleys
generate the same patterns of incomitant strabismus in
infants and children as they do in adults. Because our MRI
technique requires cooperative subjects, the youngest
subject of the current study was 17 years old. Unpublished
analyses of coronal computed tomography scans suggest
that similar heterotopic pulleys may occur at much
younger ages. These findings are consistent with the early
onset of strabismus in subjects with craniosynostosis
syndromes. It remains unknown, however, how the EOM
pulleys develop and maintain ocular alignment as the orbit
grows and matures. It should be recalled that the pulleys
are active structures, containing abundant smooth muscle
with rich innervation. 27 Pulley abnormalities therefore
could also have a neural basis. In addition, connective
tissue disorders such as Marfan's syndrome may also be
associated with pulley abnormalities (unpublished data).
In summary, abnormal EOM pulley position should be
included in the differential diagnosis of abnormal eleva-
tion or depression of the globe during adduction and
atypical incomitant strabismus. Superior mislocation of
the lateral rectus pulley of less than 2 ram, supported by
smaller mislocations of other EOM pulleys, can produce
in the fellow eye the clinical pattern of overdepression in
adduction ("superior oblique muscle overaction") and
underelevation in adduction ("inferior oblique muscle
underaction"). Inferior mislocation of the lateral rectus
pulley of less than 2 mm, with smaller mislocations of
other EOM pulleys, can produce a clinical pattern of
underdepression in adduction ("superior oblique muscle
underaction") and overelevation in adduction ("inferior
oblique muscle overaction") that masquerades as a
superior oblique palsy. Heterotopic pulleys can also
generate undesirable torsional results from muscle trans-
position surgery.
It has become conventional to describe particular
patterns of ocular motility with names that refer to
presumed causes. Terms such as "superior oblique muscle
underaction" are convenient shorthand expressions but
only if the motility disorder is truly related to impaired
superior oblique muscle contractility. Otherwise, such
expressions mislead; they beg questions of etiology and
lead to the wrong answers. We would suggest that
strabismologists gradually abandon the practice of describ-
ing motility disorder observations with diagnosis-laden
terms, except in cases where the etiology is clear.
Journal of AAPOS
Volume 2 Number I February 1998
References
1. Koornneef L. New insights in the human orbital connective tissue:
results of a new anatomical approach. Arch Ophthalrnol
1977;95:1269-73.
2. Demer JL, Miller JM, Poukens V, Vinters HV, Glasgow BJ.
Evidence for fibromuscular pulleys of the recti extraocular muscles.
Invest Ophthalmol Vis Sci 1995;36:1125-36.
3. Porter JD, Poukens V, Baker RS, Demer JL. Structure-function
correlations in the human medial rectus extraocular muscle pulleys.
Invest Ophthalmol Vis Sci 1996;37:468-72.
4. Miller JM. Functional anatomy of normal human rectus muscles.
Vision Res 1989;29:223-40.
5. Clark RA, Miller JM, Demer JD. Location and stability of rectus
musclepulleys:musclepaths as a function of gaze. Invest Ophthalmol
Vis Sci 1997;38:227-40.
6. Demer JL, Miller JM, Rosenbaum AlL. Effect of transposition
surgery on rectus muscle paths by magnetic resonance imaging.
Ophthalmology 1993;100:475-87.
7. DemerJL, MillerJM, Poukens V. Surgical implications of the recti
extraocular muscle pulleys. J Pediatr Ophthalmol Strabismus
1996;33:208-18.
8. Demer JL, Miller JM, Rosenbanm AL. Clinical correlations of
models of orbital statics. In: Scott AB, editor. Mechanics of
strabismus symposium. San Francisco: Smith-Kettlewell Eye
Research Institute; 1992. p. 141-61.
9. MillerJM, ShanaaevaI. Orbit® 1.6GazeMechanics Simulation. San
Francisco: Eidactics; 1996.
10. Robinson DA. A quantitative analysis of extraocular muscle
cooperation and squint. Invest Ophthalmol Vis Sci 1975;14:801-25.
11. Miller JM, Robinson DA. A model of the mechanics of binocular
alignment. Comput Biomed Res 1984;17:436-70.
12. Miller JM, Demer JL. Biomechanical analysis of strabismus.
Binocular Vision Eye Muscle Surg Q 1992;7:233-48.
13. yon Noorden GK. Binocular vision and ocular motifity: theory and
management of strabismus. 4th edition. St. Louis: Mosby-Year
Book; 1990.
14. Demer JL, Miller JM, Koo EY, Rosenbaum AR. Quantitative
magnetic resonance morphometry of extraocular muscles: a new
diagnostic tool in paralytic strabismus. J Pediatr Ophthalmol
Strabismus 1994;31:177-88.
15. Demer JL, Miller JM. Magnetic resonance imaging of the functional
anatomy of the superior oblique. Invest Ophthalmol Vis Sci
1995;36:906-13.
Clark et al. 25
16. Mehta AM, DemerJL. Magnetic resonance imaging of the superior
oblique muscle in superior oblique myokymia.J Pediatr Ophthalmol
Strabismus 1994;31:378-83.
17. Cheng H, Burdon MA, Shun-Shin GA, CzypionkaS. Dissociatedeye
movements in craniosynostosis: a hypothesis revived. Br J
Ophthalmol 1993;77:563-8.
18. Krzizok T, Wagner D, Kaufmann H. Elucidation of restrictive
motifity disorder in high myopia by MRI. Arch Ophthal
1997;115:1019-27.
19. Lignereux F, Duffas M, Stork LA, Gayet M, Peehereau A.
Congenital superior oblique palsy:morphological study by magnetic
resonance imaging. Ophthalmology 1996;103(suppl):178.
20. Demer JL, Miller JM, Koo EY, Rosenbaum AL, BatemanJB. True
versus masquerading superior oblique palsies: muscle mechanisms
revealed by magnetic resonance image. In: Leunerstrad G, editor.
Update on strabismus and pediatric ophthalmology. Boca Raton
(FL): CRC Press; 1995. p. 303-6.
21. Guyton DL, Weingarten PE. Sensory torsion as the cause of
primary oblique muscle overaction/underaction and A- and V-
pattern strabismus. Binocular Vision Eye Muscle Surg Q
1992;9:209-36.
22. Clark RA, Miller JM, Demer JL. Displacement of the medial
rectus pulley in superior oblique palsy. Invest Ophthalmol Vis
Sci 1998. In press.
23. Kushner BJ. Psuedo inferior oblique overactionassociatedwith Y and
V patterns. Ophthalmology 1993;98:1500-5.
24. Freedman HL, Kushner BJ. Congenital ocular aberrant
innervation--new concepts. J Pediatr Ophthalmol Strabismus
1997;34:10-6.
25. Demer JL, Miller JM. Dynamic magnetic resonance imaging of
the inferior oblique muscle in normal and strabismic subjects
[abstract]. In: Proceedings of the twenty-third annual meeting of
the American Association for Pediatric Ophthalmology and
Strabismus; 1997 Apr.; Charleston, South Carolina. Charleston:
American Association for Pediatric Ophthalmology and
Strabismus; 1997. p. 11.
26. Lockwood CB. The anatomyofthe muscles,ligaments,and fasciaeof
the orbits, including an account of the capsule of Tenon, the check
ligaments of the recti, and of the suspensory ligament of the eye. J
Anat Physiol 1885-1886;20:1-25.
27. Demer JL, Poukens V, MicevychP. Innervation of extraocularpulley
smooth muscle in monkey and human. Invest Ophthalmol Vis Sci
1997;38:1774-85.