Retinal arterial but not venous tortuosity correlates
with facioscapulohumeral muscular dystrophy
Susannah Q. Longmuir, MD,a,bKatherine D. Mathews, MD,b,cReid A. Longmuir, MD,a,e
Vinayak Joshi, MS,dRichard J. Olson, MD,a,band Michael D. Abra `moff, MD, PhDa,e
Facioscapulohumeral muscular dystrophy (FSHD) is an autosomal-dominant disease
beginning with facial and shoulder girdle weakness with variable progression. Exudative
retinal detachment, retinal vessel irregularities on fluorescein angiography, and retinal
vessel tortuosity have been found in association with FSHD.
In this retrospective study, muscle affectedness severity was rated as mild, moderate, or
severe by a neurologist masked to the retinal images. Three ophthalmologists masked to
disease severity graded the degree of arterial and venous tortuosity on a scale of 1 to 4.
An automated method estimated an index of tortuosity for arteries and veins from color
fundus photographs. Spearman rank correlation coefficients were used to describe the
relationship between retinal vessel tortuosity and disease severity.
Seven patients with an average age of 13 years (range, 7-36 years) were selected.
Correlation between the subjective tortuosity for arteries, and the severity of FSHD was
0.78 (p 5 0.039). The correlation coefficient for venous tortuosity was ?0.06 and was
not significant (p 5 0.882). The correlation coefficient between the average algorithmic
computer-generated tortuosity indices for arteries and FSHD severity was high (0.85,
p 5 0.016), but for veins it was low and not significant (0.19, p 5 0.662).
The authors of previous reports have shown retinal vascular abnormalities did not correlate
and theseverity of disease in FSHDpatients. These results suggestthetortuosity of arteries
can serve as a biomarker of severity of disease in these FSHD patients, either as determined
by human experts or by an automated method.
(J AAPOS 2010;14:240-243)
variable progression. FSHD has a highly variable clinical
presentation, ranging from patients who have a more
severe disease with onset in childhood through the usual
presentation with weakness in young adulthood to
minimally symptomatic patients in later adulthood.
Retinal vasculopathy is an established component of the
FSHD phenotype. Exudative retinal detachment, retinal
vessel irregularities on fluorescein angiography, and retinal
acioscapulohumeral muscular dystrophy (FSHD)
is an autosomal-dominant disease that begins
with facial and shoulder girdle weakness and has
vessel tortuosity have been reported in association with
retinal blood vessels in those affected by facioscapulohum-
eral muscular dystrophy have been noted in 50% to 75% of
patients.1,2In addition to the muscle weakness and retinal
abnormalities, hearing loss, supraventricular arrhythmias,
and, rarely, central nervous system manifestations also are
noted to occur in patients with FSHD. There has been
a strong link between sensorineural hearing loss and eye
complications associated with FSHD.1-5
Currently, we know that FSHD is caused by deletion of
a 3.3 kb repeat at 4q35. Severity is broadly correlated with
the number of residual 4q35 D4Z4 repeats; therefore,
patients with fewer repeats have more severe disease.
Pathophysiology is likely to involve altered regulation of
transcription, possibly of more than one gene. Microarray
analysis identified a cluster of genes with a role in vascular
biology that was up-regulated in early FSHD muscles,
leading to a hypothesis that vasculopathy may have a pri-
mary role in FSHD pathogenesis.6This study is the first
to use a novel, computer-driven analysis of arterial
tortuosity in FSHD and to show this calculated index of
Author affiliations:aDepartment of Ophthalmology;bDepartment of Pediatrics;
cDepartment of Neurology, anddDepartment of Biomedical Engineering, Universityof Iowa
Hospitals and Clinics;eVA Medical Center, Iowa City, Iowa
Presented as a poster at the 35th Annual Meeting of the American Association for
Pediatric Ophthalmology and Strabismus, San Francisco, CA, April 17-21, 2009.
Submitted October 9, 2009.
Revision accepted March 16, 2010.
Reprint requests: Susannah Longmuir, MD, University of Iowa, 200 Hawkins Drive,
Iowa City, IA 52246 (email: Susannahfirstname.lastname@example.org).
Copyright ? 2010 by the American Association for Pediatric Ophthalmology and
1091-8531/2010/$36.00 1 0
Journal of AAPOS
tortuosity correlating to disease severity and to subjective
assessment of vessel tortuosity by three different ophthal-
mologists. On the basis of these preliminary findings,
arterial tortuosity in the posterior pole could become an
additional, objective metric for severity of disease in
patients with FSHD.
2008 to October 2008 was performed, following institutional re-
view board approval. This paper conformed to the requirements
of the United States Health Insurance Portability and Account-
ability Act. Informed consent was not required by an institutional
of this study. The patient population primarily consisted of early-
onset sporadic cases. The patients included in the study had both
an examination by an ophthalmologist with fundus photographs
and an examination by a pediatric neurologist with expertise in
FSHD disease muscle affectedness severity was rated as mild,
moderate, or severe by an experienced clinician masked to the
tients were of varying ages; therefore, strict motor criteria could
notbeused forthisassignment.Ageat reportedonset ofweakness
also did not distinguish clinical severity, likely because some fam-
ilies reported onset of facial weakness long before skeletal muscle
weakness became apparent. It is also likely that some highly ob-
servant families noted problems that were quite subtle and pro-
gressed slowly. Three clinical groups could be identified,
nevertheless. Severe patients reported the use of a wheelchair
more than 50% of time, had significant functional impairment,
and wore hearing aids. The moderately affected patients used
a wheelchair only for long distances, had normal hearing, and ex-
perienced some limitations of daily activities (eg, required assis-
tance for schooling). The mildly affected patients were
itation in daily activities, and did not have any hearing loss on
All subjects underwent dilated digital fundus photography of
the right and left eyes (Carl Zeiss Meditec, Jena, Germany).
Three ophthalmologists, who were masked to the patients’ se-
verity of disease or any other identifying information, subjec-
tively assigned a tortuosity estimate based on a Likert scale of
1 to 4 for both arteries and veins from the right and left fundus
photographs. A grade of 1 indicated minimal tortuosity; 2,
minimal-to-moderate tortuosity; 3, moderate tortuosity; and
4, severe tortuosity. The three ophthalmologists were pre-
sented with both photographs of the right and left eye and
graded the 2 photographs with one index. Both the arteries
and veins were assigned independent scales of tortuosity. The
arterial and venous tortuosity scores were correlated indepen-
dently with the disease severity ratings, using Spearman rank
An automated method, currently in press, was used to
determine a quantitative, objective measure of arterial and
venous tortuosityr, termed the tortuosity index (TI).7A com-
puter program was then used to analyze the tortuosity. In
brief, 4 arteries and 4 veins from each eye were selected for
both left and right eye images from each subject; the TI was
then calculated for these arteries and veins (ATI and VTI, re-
spectively). The center line of each artery and vein was deter-
mined, and a calculation determined by the number of
changes in curvature sign, the angle of curvature of the vessel,
the ratio of arc length to the respective chord length, and total
length of vessel was performed, as explained in the following
Lc? m ? m
Where TI indicates tortuosity index; n, number of changes in
curvature sign; m, number of segments in the vessel; qi, angle
of curvature; Lci, length of the respective area; Lxi, length of
the respective chord; Lc, total length of the vessel; and *, multi-
The TIs for each artery and vein were averaged, resulting in
a single arterial TI and venous TI for each subject. The subjec-
tively assigned arterial and venous tortuosity scores, and the arte-
rial and venous tortuosity indices were correlated with the disease
severity ratings with the use of Spearman rank correlation
Of the 8 patients diagnosed with FSHD that had fundus
color photographs, 7 were included; 1 of the patients had
been diagnosed with a Coats-like bilateral retinopathy
and treated by a retinal specialist at another institution
and was therefore excluded from the study. The average
age was 13 years (range, 7-36 years); 4 of the patients
The retinal arterial and venous tortuosity scores and
the disease severity ratings are shown in Table 1. The av-
erage score for the subjective rating of the 3 ophthalmol-
ogists was then correlated to disease severity. Correlation
between the average subjective arterial tortuosity grading
by the ophthalmologists and disease severity was 0.78 (p
5 0.039). Correlation coefficient for the venous tortuos-
ity was ?0.06 and was not significant (p 5 0.882). Figures
1 and 2 show fundus photographs of patients with,
The average ATI and VTI for both eyes of the patients
are listed in Table 1. More severe disease was associated
with greater ATIs. The correlation between the average al-
gorithmic TI for arteries was 0.85 (p 5 0.016). For the
VTI, the correlation coefficient was 0.19 (p 5 0.662).
The most discordance between ATI and subjective rating
was in Patient E, who was also found to have different
average arterial indices for each eye (TI for the right eye
was 18 and for the left eye was 53) and an average TI of
36 for the arteries of both eyes.
Journal of AAPOS
Volume 14 Number 3 / June 2010Longmuir et al241
Our preliminary results strongly suggest that retinal arte-
rial,but not venous, tortuosityincreases with disease sever-
ity and could therefore serve as an easily obtainable,
objective biomarker for FSHD severity. It appears that
the arterial vessel wall connective tissue or smooth muscle
sistent with the hypothesis that vasculopathy may have
a primary role in FSHD pathogenesis. These qualitative
traits derived from the retinal vessels correlate highly
with expert determined disease severity.
ied previously relate mostly to telangiectasia, perivascular
inflammatory infiltrates, microaneurysms, and increased
permeability of the retinal blood vessels, which can occa-
sionally lead to retinal detachment and vision loss from ex-
udative retinal detachment. The authors of previous
studies have not shown any correlation with the extent of
vascular abnormalities seen on fluorescein angiography
and disease severity2; however, few early-onset patients
were studied. Fitzsimons and colleagues2showed that
mal with retinal capillary abnormalities, telangiectasis,
microaneurysms, and leakage in 56 of 75 patients with
clinical or genetic evidence of FSHD, but only 3 patients
had relevant ophthalmoscopic abnormalities in the poste-
rior pole. The findings from the angiography suggest
that retinal capillary abnormalities are an integral part of
FSHD and could be implicated in the pathogenesis.
However, Gieron and colleagues’ report8of a mother and
her 3 children highlights the tortuosity of retinal vessels
phy of FSHD. Others have described tortuosity of retinal
vessels associated with hearing loss in patients with FSHD.
Small9suggested that tortuosity of retinal vessels represents
a possible carrier state or mild manifestation of Coats
disease, which is described in association with FSHD.
Limitations of this study include the small sample size of
patients. Most of our patients traveled to a conference for
FSHD in the area and were offered complete eye examina-
tions as part of a comprehensive examination to screen for
vision-threatening complications of exudative retinopathy.
(To have access to this many early-onset FSHD patients
would require either years of observation at one center or
a multicenter trial.) Another limitation to the study is the
subjective grading of tortuosity of blood vessels, which
has already been proved to be highly variable in other dis-
eases and entities such as retinopathy of prematurity.10To
Table 1. Arterial and venous retinal scores (based on a Likert scale of 1-4) given by the 3 ophthalmologists, the average ATI, average VTI
ATI, arterial tortuosity index; VTI, venous tortuosity index.
FIG 1. Fundus photographs of the right (A) and left (B) eyes of a patient with severe FSHD. The subjective arterial tortuosity ranking was 4, 4, 4 by
the 3 observers. Subjective venous tortuosity was 1, 1, 2. The average ATI was calculated at 65. The average VTI was calculated at 9. The patient
could walk but used a wheelchair approximately 75% of the time and had documented hearing loss and used a hearing aid.
Journal of AAPOS
242Longmuir et alVolume 14 Number 3 / June 2010
decrease this bias, computer-based image analysis, which Download full-text
has been shown to produce quantifiable, objective mea-
surements, was used to confirm our findings. In our small
pilot study, the subjective assessment of an increase in
arterial tortuosity was confirmed by the generated index
of arterial tortuosity, and this index of tortuosity correlated
to the disease severity. Because retinal arterial and venous
tortuosity is part of a normal spectrum, we do not propose
that these findings should be used to diagnose FSHD, but
they may be helpful in determining which patients may
have or develop more severe disease.
The authors performed a MEDLINE search covering the
period 1950 to the present by using the terms facioscapulo-
humeral muscular dystrophy and retinal findings. Searches
were limited to articles or abstracts in English.
1. Gurwin EB, Fitzsimons RB, Sehmi KS, Bird AC. Retinal telangiecta-
sis in facioscapulohumeral muscular dystrophy with deafness. Arch
2. Fitzsimons RB, Gurwin EB, Bird AC. Retinal vascular abnormalities
in facioscapulohumeral muscular dystrophy. A general association
with genetic and therapeutic implications. Brain 1987;110:631-48.
3. Kamata T, Miyajima H, Shimizu T, Fujimoto M, Takahashi Y,
Kaneko E. Facioscapulohumeral syndrome with sensorineural hear-
ing loss and abnormality of retinal vessels. Intern Med 1993;32:
4. Padberg GW, BrouwerOF, de Keizer RJW, et al. On the significance
of retinal vascular disease and hearing loss in facioscapulohumeral
muscular dystrophy. Muscle Nerve 1995;2:73-80.
5. Taylor DA, Carrol JE, Smith ME, Johnson MO, Johnston GP,
Brooke MH. Facioscapulohumeral dystrophy associated with hearing
loss and coats syndrome. Annals of Neurology 1982;12:395-7.
6. Osborne RJ, Welle S, Venance SL, Thornton CA, Tawil R. Expres-
sion profile of FSHD supports a link between retinal vasculopathy
and muscular dystrophy. Neurology 2007;20:569-77.
7. Joshi V, Reinhardt JM, Abra `moff MD. Automated measurement of
retinal blood vessel tortuosity. Proc SPIE. In press.
8. Gieron MA, Korthals JK, Kouseff BG. Facioscapulohumeral
dystrophy with cochlear hearing loss and tortuosity of retinal vessels.
Am J Med Genet 1985;22:143-7.
9. Small RG. Coats disease and muscular dystrophy. Trans Am Acad
Ophthlmol Otolaryngol 1968;72:225-31.
10. Chiang MF, Felman R, Martinez-Perex ME, et al. Image analysis
for retinopathy of prematurity diagnosis. JAAPOS 2009;13:
FIG 2. Fundus photographs of the right (A) and left (B) eyes of a patient with mild FSHD. The subjective arterial tortuosity was 1, 1, 1 by the 3
without assistance of a walker, and did not use a wheelchair.
Journal of AAPOS
Volume 14 Number 3 / June 2010 Longmuir et al243