Major Articles

Botulinum Toxin Injection of the Inferior Oblique

Muscle for the Treatment of Superior Oblique
Muscle Palsy
Abbas Bagheri, MD, and Mohammad Eshaghi, MD

PURPOSE To evaluate botulinum toxin injection of the inferior oblique muscle for management of
superior oblique (SO) paresis.
METHODS We undertook a prospective case series of injections of the ipsilateral inferior oblique (IO)
muscle with 10-20 units of botulinum toxin type A in patients with a SO muscle
palsy/paresis of less than 2 years’ duration.
RESULTS We enrolled 16 patients (18 eyes) with a mean age of 33.7 years. The median time from
onset of symptoms until injection was 6 months. The cause of paresis was trauma for 81%
of the patients. The mean hypertropia decreased from 6.4⌬ to 1.9⌬ at 6 months after
treatment. Other clinical measures improved: mean IO overaction from ⫹ 1.7 to ⫹ 0.6,
mean SO underaction from –1.5 to – 0.4, mean subjective torsion from 9.3° to 0.4°, and
mean head tilt from 8.4° to 1.1°.
CONCLUSIONS Botulinum toxin injection to IO muscle reduces the symptoms of SO paresis while patients
are waiting for recovery. ( J AAPOS 2006;10:385-388)

T
he superior oblique (SO) muscle frequently is im- Informed consent was obtained from each patient before the
plicated in paralytic strabismus.1,2 The acquired procedure. Eligible patients had an isolated unilateral or bilateral
form causes diplopia, often with an anomalous acquired SO palsy/paresis of less than 2 years’ duration. In
head posture, which may make normal activities difficult. addition, they had no history of eye surgery, no sign of contrac-
Most patients are followed for 6 months to a year until the ture in the ipsilateral superior rectus muscle, and were available
signs stabilize and an individualized treatment decision can to complete a 6-month outcome examination.
be made for these patients. Temporary measures generally SO palsy/paresis (hereafter termed paresis) was confirmed by
include occlusion and prism correction. However, obser- the Bielschowsky 3-step test. Prism measurements of the devia-
vation may be associated with muscle contracture, which tion were performed in primary position and all diagnostic gazes.
could result in a less satisfactory long-term result or re- Unilateral cases were measured with the sound eye fixating and
quire more surgery to correct the strabismus. the prism placed before the paretic eye. Subjective torsion was
Botulinum toxin type A (BTA) has been used for treat- measured by double Maddox-Rod test, objective torsion evalu-
ment of oculomotor and abducens nerve pareses, and some ated by fundus photography, and abnormal head posture mea-
authors have suggested that it may decrease the need for sured by torticulometer and head deviometer as described by
surgery.1,2 Studies using this toxin for the treatment of Bagheri et al.6
trochlear nerve paresis are limited.3-5 This study was de- Grading of IO overaction was performed by dividing the
signed to evaluate the results of injection of BTA into the cornea vertically into 4 segments. With the affected eye directed
inferior oblique (IO) muscle for the treatment of SO up in adduction, we graded the IO overaction by the number of
palsy/paresis. segments of the cornea that were above the reflex in the fellow
eye. Grading of SO underaction was performed by assessing the
Materials and Methods number of corneal portions the affected eye was above the fellow
This was a prospective uncontrolled treatment trial. The study eye.
was approved by an institutional research oversight committee.
Preliminary Study
This work was undertaken to develop an injection technique for
Author affiliations: Department of Ophthalmology, Labbafinejad Hospital, Shahid
Beheshti University, Ophthalmic Research Center, Tehran, Iran the IO muscle without electromyogram (EMG) or conjunctival
Submitted January 9, 2006. surgery. Eligible patients were those about to undergo IO sur-
Revision accepted June 13, 2006. gery. The IO muscle was isolated with an inferotemporal fornix
Reprint requests: Abbas Bagheri, MD, Labbafinejad Hospital, Ophthalmic Research
Center, Boostan 9 str., Pasdaran Avenue, Tehran, Iran 16666 (email: abbasbagheri@ conjunctival incision. The distance from the anterior border of
yahoo.com). IO muscle to the limbus was measured with a curved caliper with
Copyright © 2006 by the American Association for Pediatric Ophthalmology and the eye rotated superonasally. We studied 35 eyes (19 right, 16
Strabismus.
1091-8531/2006/$35.00 ⫹ 0 left) of 23 patients with a mean age of 23 years (range, 5 to 63).
doi:10.1016/j.jaapos.2006.06.014 The mean distance from the limbus to the anterior border of the

Journal of AAPOS 385

386 Bagheri and Eshaghi Volume 10 Number 5 October 2006

FIG 1. Illustration of the distance from limbus to the anterior border of IO

muscle in the inferotemporal quadrant (illustration by Dr. Ramin Saheb-
FIG 2. Schematic figure of technique of injection of toxin in the IO muscle
ghalam).
(illustration by Dr. Ramin Sahebghalam).
IO muscle was 13.6 ⫾ 0.9 mm (range, 12.0 to 16.0). The muscle
width has been reported to be 5-14 mm (average, 9.0 mm).7 For data analysis, dichotomous outcomes were assessed with
Combining these findings suggests that the needle tip should be the Wilcoxon rank-sum test. A paired t-test was used to analyze
positioned 20 mm from the limbus to be in the belly of the IO continuous data.
muscle (Figure 1).
Results
Injection Technique
The study enrolled 18 eyes of 16 patients from March
Botulinum toxin type A (BTA; Dysport, Ipsen, Paris, France) was
2003 to February 2004 with SO palsy/paresis. The mean
injected transconjunctivally with topical tetracaine (%0.5) in the
age of the patients was 33.7 ⫾ 17.2 years (range, 6 to 72
inferior temporal quadrant without EMG guidance and without
years). There were 14 men (87%) and 2 women (13%).
using a speculum or forceps. A 1 mL syringe with a 1 inch, 23
The mean duration of paresis was 6 ⫾ 6.4 months (range,
gauge needle was used. The patient was instructed to look supe-
1 to 18 months) which included 9 eyes (50%) with less
riorly in adduction with the affected eye. With the bevel of the
than 6 months of disease and 9 eyes (50%) with more than
needle toward the sclera, the needle was introduced into the
6 months of disease. The cause was trauma in 13 patients
conjunctiva 10 mm from the limbus and advanced 10 mm in-
(81%), vascular in 2 (13%), and viral in 1 (6%). The SO
ferotemporally to reach the site of belly of the IO muscle. The
palsy was unilateral in 14 patients (88%) whereas both eyes
body of the syringe was then rotated approximately 30° upward
were affected in 2 patients (12%). Ten (62%) eyes were
to move the needle away from sclera and toward the belly of the
right eyes and 4 (25%) left. Binocular diplopia was present
muscle (Figure 2). Injections included 10 or 20 units of Dysport
in all patients. A contralateral head tilt was present in 11
reconstituted in 0.1 mL of normal saline solution. The Dysport
patients (56%) and averaged 12.3 ⫾ 3.4° (range, 10-20°). A
has one quarter the potency of Botox® (Allergan, Irvine, CA);
horizontal face turn was present in 7 patients (44%) and
thus, 10-20 units of Dysport is equal to 2.5-5 units of Botox.
averaged 16.4 ⫾ 8.5° (range, 10 to 30°). Chin-down pos-
The patients were monitored in the office for 30 minutes for
ture was present in 3 patients (19%), each about 10°. SO
symptoms and signs of scleral perforation, including floaters,
underaction was present in 16 eyes (89%) with a severity of
severe ocular pain, decreased vision, subconjunctival hemor-
[minus]1 to [minus]2.
rhage, and orbital hematoma. Follow-up examinations were per-
The BTA dose was 20 units in 14 eyes and 10 units in 4
formed at 1 week, 1 month, 3 months, and 6 months after the
eyes. Diplopia improved in 15 patients (94%) within a few
injection. These examinations were performed by the investiga-
days of injection. Six months after the injection 10 eyes
tors. The outcome was defined as good, fair, or poor.
(56%) did not have underaction of SO muscle, and 8 eyes
● Good result: orthotropic without bothersome symptoms (44%) had [minus]1 to [minus]2 underaction. The change
and surgery not planned. from baseline to 6 months was significant ( p ⫽ 0.001).
● Fair result: a decrease in clinical signs of more than 50%, Mean SO underaction before injection was 1.5 ⫾ 0.7, and
but the patient had some tolerable symptoms. 6 months after treatment was 0.4 ⫾ 0.6 (Figure 3A).
● Poor result: No change in symptoms, with or without a IO overaction was present before injection in 16 eyes
change in the deviation. (88.9%) ranging from 1⫹ to 3⫹ whereas no IO overaction

Journal of AAPOS
Volume 10 Number 5 October 2006 Bagheri and Eshaghi 387

FIG 3. Severity of oblique muscle dysfunction before and after injection. A, SO muscle underaction; B, IO muscle overaction.

FIG 4. Severity of subjective torsion before and after injection.

FIG 6. Abnormal head postures before and after injection.

jection was associated with an increased likelihood of a

poor outcome; however, there was no statistically signifi-
cant relationship between specificities noted in the table
and quality of results.
Only a few side effects of treatment were noted. A mild
paresis of the inferior rectus muscle developed in 2 pa-
tients who received 20 units of BTA. Symptoms were
FIG 5. Primary position hypertropia before and after injection. evident in 2-4 days. Each resolved within one month.
There were no cases of blepharoptosis.
was present in 2 eyes (11%). Six months after treatment,
12 eyes (66.7%) did not have IO overaction, whereas 6 Discussion
eyes (33.3%) had IO overaction from 1⫹ to 2⫹ ( p ⫽ We found IO chemodenervation to rapidly restore a more
0.003). Mean IO overaction before treatment was 1.7⫹ ⫾ normal binocular status in nearly all of our patients. Most
0.9 and 6 months after treatment it was 0.6⫹ ⫾ 0.7 (Figure patients had substantial relief of symptoms as well. We do
3B). not know from this pilot study whether this benefit led to
Subjective torsion was present in 13 patients (82%) a long-term improvement in signs or a reduction in the
before injection. The average torsion before injection was need for strabismus surgery compared with the natural
9.3 ⫾ 6.9° (range, 4° to 30°), which decreased to 0.4 ⫾ 1.4° history of SO paresis. A prospective randomized treatment
6 months after treatment. ( p ⫽ 0.001; Figure 4). The mean trial is needed to test this hypothesis
hypertropia before treatment was 6.4 ⫾ 4.9⌬, which de- The injection technique used in this study was easily
creased to 1.9 ⫾ 4.5 ␳ at 6 months after treatment. ( p ⫽ learned. We injected the IO muscle without EMG guid-
0.005; Figure 5). Mean abnormal head posture before ance obtaining the desired paralytic effect in every case.
treatment was 8.4 ⫾ 6.5°, which decreased to 1.1 ⫾ 2.5° 6 Three previous studies of IO injection were performed
months after treatment ( p ⫽ 0.001; Figure 6). with EMG guidance.3-5 Lozano-Pratt and coworkers had
After 6 months, a study-defined good outcome was to repeat the injection in one case (11%) because the
found in 10 (62.5%), fair in 4 (25%), and poor in 2 muscle could not be found with EMG.3 Similarly, Buon-
(12.5%). There was very little change in these proportions santi et al4 could not find the IO with EMG in one case.
between 1 month and 6 months after the injection. The Injection without EMG guidance of the medial rectus
patients were subdivided by 6-month outcome group as- muscle has been reported for the abducens nerve paresis.8
signment at first visit and the baseline characteristics of It has been suggested that it is not necessary to inject the
these subgroups analyzed (Table 1). Duration before in- belly of the muscle to produce paralysis but that injection

Journal of AAPOS
388 Bagheri and Eshaghi Volume 10 Number 5 October 2006

Table 1. Baseline characteristics grouped by outcome

Duration of Primary
paresis deviation SOUA Subjective Head posture
N Age ( yr) (months) (⌬) (⫺) IOOA (⫹) torsion (°) (°)
Good 10 39.6 ⫾ 18.1 4.7 ⫾ 3.5 6.4 ⫾ 5.4 1.4 ⫾ 0.8 1.6 ⫾ 1.1 9.0 ⫾ 8.5 7.0 ⫾ 6.3
Fair 4 21.5 ⫾ 13.2 6.5 ⫾ 4.1 4.3 ⫾ 3.5 1.5 ⫾ 0.6 1.8 ⫾ 0.5 5.3 ⫾ 5 13.8 ⫾ 4.8
Poor 2 29.0 ⫾ 1.4 12.0 ⫾ 8.5 11.0 ⫾ 1.4 2.0 ⫾ 0 2.0 ⫾ 0 5.0 ⫾ 0 5.0 ⫾ 7.1
SOUA: superior oblique underaction; IOOA: inferior oblique overaction.

in the surrounding connective tissues near the muscle is Acknowledgments

sufficient.9 In 2 of 18 eyes (11%) toxin did spread to the
We thank Dr. Michael X. Repka for his help in preparing the manu-
adjacent IR muscle. In both cases, we had injected 20 units
script.
rather than 10 units. Buonsanti et al4 found in 5 cases
(27%) that the toxin spread to the lateral and IR muscles.
They found that higher doses increased the risk of the References
toxin spreading to other muscles. By injecting 10 units of
1. Liesegang TJ, Deutsch TA, Gilbert Grand M. American Academy of
Botox (approximately equal to 40 units of Dysport), the
Ophthalmology Basic and Clinical Science course (section 6) Pediatric
risk reaches 50%. They therefore recommended using Ophthalmology and Strabismus, 2002-2003. American Academy of
2.5-5 units of Botox, which is equal to 10-20 units of Ophthalmology, San Francisco, 2002;115-118.
Dysport used in this study.4 2. Rosenbaum AL, Santiago AP. Clinical strabismus management. Phil-
Subgroup analyses found the chance of a good outcome adelphia: W.B. Saunders Company; 1999.
at 6 months diminished with increased duration and with 3. Lozano-Pratt A, Estanol B.Treatment of acute paralysis of the fourth
cranial never by botulinum toxin A chemodenervation. Binocular
increased magnitude of IO overaction and SO underac-
Vision Strabismus Q 1994;9:155-8.
tion. These findings differ from Buonsanti et al, 4 who 4. Buonsanti JL, Riverosanchez-Covisa ME, Scarfone H, Lynch, J. Botu-
showed that when the IO overaction increases, the re- lium toxin chemodenervation of the inferior oblique muscle for
sponse to BTA injection increases. chronic and acute IV nerve palsies: results in 15 cases. Binocular
This study is limited by the lack of an untreated com- Vision Strabismus Q 1996;11:119-24.
parison group and non-naïve observers. We cannot deter- 5. Merino P, Gomez DE, Liano P, et al L. Bilateral superior oblique
palsy and botulinum toxin. Arch Soc Esp Oftalmol 2004;79:119-24.
mine how much of the long-term improvement was the
6. Bagheri A, Farahi A, Yazdani S. The effect of bilateral horizontal
result of natural history rather than that of the chemode- rectus recession on visual acuity, ocular deviation or head posture in
nervation. Its benefit also must be compared with the cost patients with nystagmus. J AAPOS 2005;9:433-7.
of the vial used in each patient and the potential compli- 7. Von Noorden GK, Campos EC. Paralytic strabismus. In Von Noor-
cations specially globe perforation and retrobulbar hem- den GK, Campos EC, editors. Binocular vision and ocular motility.
orrhage. Finally, the sample size is too small to make 6th ed. St. Louis: CV Mosby; 2002. p. 414-58.
definitive conclusions about the effect of baseline charac- 8. Benabent EC, Hermaosa PG, Maria Arazola T, Alio Y, Sans JL.
Botulinum toxin injection without elecrtomyographic assistance. J Pe-
teristics on outcome. In summary, the injection of BTA diatr Ophthalmol Strabismus 2002;39:231-4.
into the IO muscle during the acute phase of SO palsy/ 9. Kao LY, Chao AN. Subtenon injection of botulinum toxin for treat-
paresis safely and rapidly reduces the symptoms and signs ment of traumatic sixth nerve palsy. J Pediatr Ophthalmol Strabismus
of the abnormality. 2003;40:27-30.

Journal of AAPOS