EXTRAOCULAR MUSCLE PROBLEMS IN

THYROID EYE DISEASE

PETER FELLSl, LINDA KOUSOULIDESI, ANASTASIA PAPPAI•2,

PETER MUNR02 and JOANNA LAWSONl
London

SUMMARY achieve a diagnosis was necessary before special scanning

In this paper methods of visualisation of the extraocular techniques became available. The huge muscles in a
muscle changes in thyroid eye disease are discussed. The patient with 13 mm of unilateral proptosis but no other
histopathology of extraocular muscle biopsies has been mass in any quadrant confirmed the diagnosis of thyroid
studied by both light and electron microscopy to show the eye disease and also established the fact that a difference
type of cellular infiltration and the amorphous material of more than 6 mm proptosis between the two eyes did not
in the extracellular matrix. A series of Questions to which mean some cause other than thyroid eye disease as had
answers have not yet been found concerning thyroid eye previously been accepted.
disease are posed which may help to direct new research Computed tomography (eT) scans I at first gave axial
projects. Finally, in the last part of the paper, the surgical views of the orbits; then direct coronal slices could be
results in a series of 41 patients having ocular muscle sur­ taken with the patient's neck hyperextended in the large­
gery for diplopia and/or compensatory head postures due
diameter port of the whole body scanner (which was not
to thyroid eye disease are described. The conclusions
possible with the dedicated head and neck scanners).
drawn from these results are that one should maintain the
Reformatted coronal views are used now, together with
patient euthyroid, establish by orthoptic measurements
reformatted views along the optic nerve axis. Proptosis is
that the ocular movements have been stable for at least 6
the first evidence of auto-decompression, followed by out­
months, treat by recessing tight muscles using adjustable
sutures, and aim to undercorrect the vertical deviation at
ward bowing of the medial wall encroaching into the eth­
the time of adjustment. moidal space which makes the 'Coca-cola bottle' sign
when it occurs bilaterally.
Once post-mortem examinations of the orbits of patients Magnetic resonance imaging (MRI) can give even more
with thyroid eye disease had revealed the enormous swell­ information about the muscle enlargement without any
ing of the extraocular muscles it was gradually realised radiation risk. The enviable ease of repeated views to
that the muscular changes may be responsible for many of assess response to therapy is offset by their high cost in
the symptoms and signs in this condition. The first part of
some parts of the country.
this paper will review ways in which the extraocular
Ultrasonography in B-scan mode can show the size of
muscles can be investigated. The second part will pose
some of the rectus muscles in the anterior part of their
some of the many Questions that need to be answered in
course. A-scan mode has been used to assess ocular
this condition. Finally, the results of surgical intervention
muscle reflectivity, with echogenicity lower than 40%
to correct some of the consequences of these pathological
suggesting that a better response to therapy with immuno­
changes in the muscles are discussed.
suppressives may be expected.2
INVESTIGATION OF THE EXTRAOCULAR
MUSCLES Histopathology of Extraocular Muscle Biopsies

Visualisation of Extraocular Muscle Changes

Extraocular muscle biopsies from patients with thyroid
eye disease were examined by means of immunohistoche­
Surgical exploration of gross unilateral proptosis to
mistry and light and electron microscopy. The biopsies
From: 'Moorfields Eye Hospital. London; 2Institute of were taken under local anaesthesia from the belly of the
Ophthalmology. London, UK.
Correspondence to: Peter Fells, FRCS. Moorfields Eye Hospital, City involved recti muscles, often during the early stages of the
Road. London EC1V 2PD. UK. disease before ocular rotations were reduced and before

Eye (1994) 8, 497-505 © 1994 Royal College of Ophthalmologists

498 P. FELLS ET AL.

Fig. 1. Light micrograph of transversely cut extraocular

muscle tissue from a patient with thyroid eye disease. Staining
with toluidine hlue/hasic fuchsin on a resin section shows Fig. 2. Transmission electron micrograph illustrating a peri­
increased interfibrillar spaces, fatty cell infiltration (arrow) as venular inflammatory cell infiltrate in association with a longi­
well as increased numhers of collagenfihres denoted by the red tudinally cut extraocular muscle fihre of a thyroid eye disease.
colour. (x750) extraocular muscle hiopsy. Scale bar represents 5 J1-m (x 2 200)

Fig. 4. Light micrograph of immunoperoxidase staining for

Fig. 3. Light micrograph of immunoperoxidase staining for HLA class II (DR) on a cryostat section of extraocular muscle
intracellular adhesion molecule 1 (/CAM-I) on a cryostat section tissue from a patient with thyroid eye disease. Positively stained
of extraocular muscle tissue from a patient with thyroid eye cells and their processes are located in the endomysium and per­
disease. /CAM-I expression is seen as a brown precipitate. (x750) imysium (hrown precipitate). (x476)

immunosuppressive therapy was started. The inferior and and fat cells, the presence of which adversely affects the
medial recti were the most commonly biopsied muscles. contractile properties of the affected muscle.
Microscopic examination demonstrates either focal or Lymphocytic infiltration coincides with upregulation of
diffuse infiltration of the spaces between the extraocular adhesion molecules3 on the surface of both lymphocytes
muscle fibres with mononuclear inflammatory cells and vascular endothelium (Fig. 3). Increased expression
including lymphocytes, macrophages and mast cells of HL A class II molecules can be detected by means of
(Figs. 1, 2). The interfibrillar spaces were enlarged and immunohistochemistry on the surface of the cells from the
contained an amorphous material in the extracellular endomysium and perimysium, probably fibroblasts
matrix. This amorphous material has been shown immu­ (Fig. 4). Some of the infiltrating lymphocytes are acti­
nohistochemically to have hyaluronic acid as a major vated and produce lymphokines which either attract other
component, using a recently developed monoclonal anti­ inflammatory cells into the area or stimulate fibroblasts.4-6
body against hyaluronic acid. Stimulated fibroblasts in tum produce increased quantities
Infiltration of the endomysial space by lymphocytes, of glycosaminoglycans (linear polysaccharide molecules)
macrophages and neutrophils is observed during the early which osmotically attract water causing interstitial
stages of the disease. These cells decrease in numbers as oedema.7 The fibroblasts are also more numerous in the
the disease bums itself out and are replaced by collagen connective tissue. Muscle fibres themselves are thought
THYROID EXTRAOCULAR MUSCLE PROBLEMS 499

not to be affected and to be similar to control fibres, but triction than expected. Conversely in some patients with a
morphometric measurements of muscle fibre area and dia­ positive traction test, i.e. decreased movement, rotation
meters have yet to be completed. may not return to normal when the appropriate rectus
muscle has been fully detached during surgery.
Muscle Activity Analysis Botulinum injections within a year of onset of restricted
Muscle activity can be analysed by means of (l) length­ movements can give much improved movement.s This
tension curves of extraocular muscles, (2) electromyog­ cannot be affecting fibrous tissue but must be working on
raphy (EMG) and (3) intraocular pressure changes in cer­ the muscles to relax the active contraction that Simonsz
tain gaze directions. and Kommere1l9 have shown. Unfortunately the rapid
change from hypo- to hypertropia after inferior rectus
Orthoptic Testing toxin upsets many patients, so we no longer use it.
After surgical orbital decompression,!O which for many
Orthoptic testing can be done using (1) Hess charts, (2)
surgeons means using the ethmoidal space and the antral
field of binocular single vision (BSV) and (3) uniocular
sinus medial to the infraorbital nerve, patients often show
fields of fixation. We first used this last method in 1969 but
increased esotropia, usually with an A-pattern. When the
abandoned it because the results were inconsistent and
orbital resistance has been reduced the tight medial recti
provided no more useful information than the Hess charts
may pull the eyes into esotropia. Esotropia may be
and BSV fields elicited. We tried again in 1991/2 but with
increased by tight inferior recti too. The A-pattern results
similar lack of success and no longer use that method.
from disinsertion of the inferior oblique muscles when the
The methods we find most useful in clinical practice are
medial part of orbital floor is removed, and the reduced
CT scans, serial Hess charts and fields of BSV.
proptosis also encourages increased superior oblique
action.
QUESTIONS IN THYROID EYE DISEASE
The timing and sequence of surgery to the extraocular
Some major questions in thyroid eye disease are: muscles is most important.!! If orbital decompression is
Are the signs of optic neuropathy due to direct com­ indicated it must be done first. Even when decompression
pression of the optic nerve by enlarged muscles at the orbi­ is not required in the absence of optic neuropathy or severe
tal apex? corneal exposure, if proptosis is greater than 26 mm then
How many signs of thyroid eye disease can be attributed to recessing tight recti can make the proptosis worse. Hence
orbital venous stasis from vein compression (a thesis proptosis of 26 mm and more should be reduced before
developed by Feldon)? ocular muscle surgery. Next, ocular movement stability is
How can clinically unilateral proptosis occur? essential before trying to correct diplopia. In the past this
Why are not all the extraocular muscles enlarged equally? would mean waiting 2 years before ocular muscle surgery.
We use stable ocular measurements on the Hess chart and
Are the oblique muscles involved?
BSV fields for at least 6 months as our criteria. Continuing
How can the proptosis of the second eye be delayed, some­ active orbital inflammation needs to be identified and
times for many years, and still be due to thyroid eye avoided. Mourits developed an index of clinical activity
disease? using the long-established criteria for inflammation of
Why do some patients have slight proptosis and signs of pain, redness, swelling and impaired function. We have
severe optic nerve compression? found this difficult to use, in particular when attempting to
Why do some patients have marked proptosis and no signs characterise the different types of orbital pain, although
of optic nerve compression? the changes in oedema are helpful. Other objective
How do some patients have marked proptosis and no signs methods of assessing active inflammation include using
of extraocular muscle enlargement? ultrasonography in A-scan and B-scan modes, and the dif­
Are the reduced ocular rotations always due to tight ferent T spins in MRI scanning. Kahaly and, independen­
muscles restricting movement? tly, one of the present authors (A.P.) have demonstrated
increased urinary glycosaminoglycans (mucopolysaccha­
Does ocular muscle paresis occur as well?
rides) in the active phase of orbital inflammation.
When is the right time to operate on the extraocular
muscles? SURGICAL RESULTS
How does botulinum toxin injection work in thyroid eye A series of 41 patients underwent ocular muscle surgery
disease? for diplopia and/or compensatory head posture due to thy­
How can botulinum toxin give long-lasting improvement? roid eye disease. Twenty-eight were female and 13 male,
of whom 25 were cigarette smokers and 7 had other auto­
We have partial answers to a few of these questions. immune diseases such as pernicious anaemia and vitiligo.
Reduced ocular movements are normally due to tethering Thirty patients were hyperthyroid initially, 6 hypothyroid
by tight rectus muscles, the inferior and medial recti being and 5 euthyroid, although all had been rendered euthyroid
most often involved. However, careful performance of the by the time of their strabismus surgery. Thirteen had been
traction test under general anaesthetic may reveal less res- treated by radioactive iodine, 2 by partial thyroidectomy
500 P. FELLS ET AL.

-------- ---- ---- ----

Group 1: Excellent Outcome

60 •
•

50 •
• pre-op

I
c •

. 2 40 • • o post-op

I
i
. • •
final
:; • •

CIl 30 •

0 I

l 1 1 6-ti
'iij 20

1
u •

t:
CIl 10
> I
�.. .
0 � Q
.+.-+-.
'" r-- en ("') Lt'l rt.

-10

patient no.
- --- -- --

(a)

Group 1: Excellent Outcome

50 •

40
c

1 i
. 2
i
.
30
:;
CIl
20

i
0 • • pre-op

.----�-+-�.--+---r----'
r----1 1
5c 10 , � • o post-op
0
N
.>: 0 -� I- I I
• final
0 I
J:
-10
2 .� � 5 6
u

-20

patient no.

_._ .. _------------- ----- ------'

(b)

Group 1: Excellent Outcome

90 �----.

80

70
"
:.e
60
III
! 50
III
>
U) 40
ID
30

20

10

patient no.

I [J pre-op post-op. final I

(c)
Fig.S. Results o/the excellent outcome group (group I) showing the improvement in the ver­
tical deviation, horizontal deviation and area of hinocular single I'ision./i)llowing surgery.
THYROID EXTRAOCULAR MUSCLE PROBLEMS 501

Group 2: Good Outcome

60 •

50 •
•
c 40
0
•
•
:;

� I�rI
30
• [ •

.[
'S;

�
GI 20 •

I i
0 •

iii 10 [
u �
'f 0 � I� • pre-op
GI .
. ..,. � <0 � �. <» 0

1
> -10 (' [] post-op
-20 � • final
-3 0

patient no.

(a)

Group 2: Good Outcome

40 •

• •

30 • •
c
0
:; 20
'S;
GI
. >
0 0 .
10 0
jc • pre-op
0 • [] post-op
0 . .

�
,!:II
...
0 a 4 5 6 • final
::E: -10

-20

patient no.

(b)

Group 2: Good Outcome

90 �-----'
80
70
III 60
t
'W 50
>
� 40
*' 30
20
10
o

patient no.

10 pre-op post-op. final I

(c)
Fig. 6. Results of the good outcome group (group 2 ) showing the improvement in the ver­
tical deviation, horizontal deviation and area of binocular single vision following surgery.
502 P. FELLS ET AL.

and some were still on carbimazole. Both primary and were recessed: the inferior rectus in 34 procedures, medial
secondary hypothyroidism were treated by T4• Fourteen rectus in 14 and the superior rectus in 9.
patients had received oral prednisolone, 8 radiotherapy Twenty patients had excellent results of BSV, extending
and 8 surgical decompression; 7 had had botulinum toxin downwards more than 30° without the help of prisms or a
injections. compensatory head posture. Fourteen patients had good
Thirty-two patients had one strabismus operation, 8 had results where the same area of BSV needed a small com­
two operations and 1 had four operations. Of these oper­ pensatory head posture or not more than 6 dioptres of
ations 32 were vertical procedures, 16 combined vertical prism (Fig. 6). Four satisfactory and 2 poor results showed
and horizontal, and 4 horizontal procedures (Fig. 5). smaller areas of BSV with worsening head postures and
Adjustable sutures were used in 47 operations but in 24 stronger prisms.
cases no adjustment was necessary and the suture was Muscle surgery in thyroid eye disease is almost unique
simply tied off. In all cases except 2 the tight muscles only in often requiring adjacent rectus muscles to be operated

tomp. tomp.

(a)

tomp. temp.

(b)
Fig. 7. Hess charts of patient 1 hefore (a) and 1 day after right superior rectus recession (h). (Continues).
THYROID EXTRAOCULAR MUSCLE PROBLEMS 503

on at the same time, usually inferior rectus and medial rec­ The powerful traction necessary to pull the globe into a
tus. In 1978 one of us (P.F.) first reported12 that this could position where muscle surgery can be performed can
cause anterior segment ischaemia in the lower nasal quad­ cause the intraocular pressure to rise to 50-70 mmHg.
rant with slight pupillary irregularity and patchy loss of Such traction can be sustained safely for short periods
posterior iris pigment in this quadrant. Recently P.F. has, only. Occasionally the tightness is such that it seems
with Jane Olver, been using a vessel-sparing technique for impossible even to insert sutures into the muscle. The
recessions of (usually) the inferior rectus but leaving one temptation to do a free tenotomy should be resisted since
or more of its anterior ciliary vessels intact. Naturally ana­ this will invariably be followed by overcorrection. Perse­
tomical preservation has to have post-operative patency verance virtually always allows recession to be done.
demonstrated by fluorescein angiography. Sometimes the For the best long-term results a first day post-operative
muscles are so tight that the globe cannot be rotated into a undercorrection of vertical deviation must be achieved.
position which permits vessel-sparing surgery. Patient I with right hypertropia of 22 prism dioptres

temp.

(,�czn bczfo� Icztt eye

(c)

tomp. -+ noool tomp.

(d)
Fig. 7 (continued). Five months later the right hypertropia of 2 2 prism dioptres had overcorrected to 30 prism dioptres of left hyper­
tropia (c). Finally, right inferior rectus recession gave all excellent result (d).
504 P. FELLS ET AL.

t�mp. no�ol

(a)

Fig. 8. Hess chart of patient 2 with a large esotropia and right

hypertropia (a), After hilateral medial and inferior rectus reces­
sions, using adjustahle sutures on the two muscles on the right
side, she had an excellent result and regained a BSVfield of71%
oj'normal (h, c),

(c)
THYROID EXTRAOCULAR MUSCLE PROBLEMS 505

(Fig. 7a) responded to a right superior rectus recession of Institute of Ophthalmology. Dr Anastasia Pappa's research is
supported by a grant from the Royal National Institute for the
6 mm by being straight in the primary position (Fig. 7b).
Blind. I am indebted as always to Jenny Rignold for preparing
Two weeks later she had overcorrected to left hypertropia the manuscript.
of 22 prism dioptres, increasing over 5 months to 30 prism
dioptres (Fig. 7c). This responded well to right inferior Key words: Muscles, Surgical correction, Thyroid eye disease.
rectus recession of 6 mm, together with right medial rectus
recession of 4 mm to give a final, lasting result of right
hypertropia of 3 prism dioptres (Fig. 7d). REFERENCES
Patient 2 with 35 prism dioptres of esotropia and 8 l. Trokel SL, Jakobicc FA. Correlation of CT scanning and
prism dioptres of right hypertropia (Fig. 8a) underwent pathologic features of ophthalmic Graves' disease. Ophthal­
mology 1981;88:553-64.
four muscle recessions at one operation. The right inferior
2. Ossoinig KC. Ultrasonic diagnosis of Graves' ophthal­
rectus was recessed 6 mm and the right medial rectus 5 mology. In: Gorman CA, Waller RR, Dyer JA, editors. The
mm, both on slides, and the left inferior rectus and left eye and orbit in thyroid disease. New York: Raven Press,
medial rectus were both recessed 4 mm using fixed sutures 1984: 185-211.
(Fig. 8b). From having no area of BSV the operation gave 3. Pappa, A. Fells P, Lightman S. Adhesion molecules in the
immunopathogenesis of thyroid-associated ophthalmo­
a field of BSV of 71% of normal (Fig. 8c).
pathy. In: Kaufmann H, editor. Transactions of the XXth
One of the major complications of recessing very tight Meeting of the European Strabismological Association,
inferior recti is the risk of significant post-operative lower Brussels 1992: 365-6.
lid retraction measured as the height of sclera visible 4. Bahn RS, Heufelder AE. Orbital connective tissue in endo­
between the 6 o'clock limbus and the lid margin. Dissec­ crine ophthalmopathy. In: Kahaly, G, editor. Endocrine oph­
thalmopathy: molecular, immunological and clinical
ting off the lower lid retractors from the inferior rectus
aspects. Basel: Karger, 1993: 47-57.
muscle sheath as far back as possible, i.e. 15-16 mm from 5. Schuler M, Hansen C, Winand R, Slaer C, Otto E, Beyer J,
the inferior rectus insertion, helps to reduce but does not Kahaly G. Urinary and plasma glycosaminoglycans in endo­
abolish this effect. Other surgeons are trying different crine ophthalmopathy. In: Kahaly G, editor. Endocrine oph­
methods but none is universally successful yet. Our thalmopathy: molecular, immunological and clinical
aspects. Basel: Karger, 1993: 58-67.
figures for 36 inferior rectus recessions in 30 patients (25
6. Pappa A, Jackson P, Munro P, Fells P, Lightman S. Glycosa­
procedures using adjustable sutures, 12 of which required minoglycans in the pathogenesis of thyroid-associated oph­
adjusting), were as follows. The inferior rectus recessions thalmopathy. In: Kaufman H, editor. Transactions of the
were between 3 and 8 mm. Dividing them into two groups XXIst Meeting of the European Strabismological Associ­
of small inferior rectus recessions of 3-5 mm and large ation, Salzburg, 1993: 359-63.
7. Bahn RS, Hcufelder AE. Pathogenesis of Graves' ophthal­
recessions of 5.5-8 mm the pre-operative lower lid retrac­
mopathy. N Engl J Med 1993;329:1468-75.
tion was 0.29 mm and 0.20 mm respectively. After surgery 8. Lyons CJ, Vickers SF, Lee JP. Botulinum toxin therapy in
these figures were 0.86 mm (range 0-3 mm) and 2.0 mm dysthyroid strabismus. Eye 1990;4:538--40.
(range 0-4 mm). This emphasises the importance of the 9. Simonsz HJ, Kommerell G. In Graves' disease, increased
correct order for surgery in thyroid eye disease, strabismus muscle tension and reduced elasticity of affected muscles is
primarily caused by active muscle contraction. Neuro-Oph­
surgery to align the eyes first and finally lid surgery to
thalmol 1989;9:243-6.
match the lids to the globes. 10. Fells P. T hyroid-associated eye disease: clinical manage­
The recommendations for successful surgery in this dif­ ment. Lancet 1991;338:29-32.
ficult condition are: make and maintain the patient euthy­ II. Fells P, McCarry B, Aylward Gw. Ocular muscle surgery in
roid, have demonstrably stable ocular movements for at thyroid eye disease. Orbit 1992;11:169-75.
12. Fells P, Marsh RJ. Anterior segment ischaemia following
least 6 months, recess tight muscles using adjustable
surgery on two rectus muscles. In: Reinecke RD, editor.
sutures and aim to undercorrect the vertical deviation.
Strabismus. Proceedings of the third meeting of the Inter­
The immunohistochemistry and electron microscopy were both national Strabismological Association, Kyoto. New York:
carried out in Professor Susan Lightman's department at the Grune & Stratton, 1978:375-80.