You are on page 1of 7

ecent advances in color Doppler imaging

techniques permit percutaneous visualiza-


tion and hemodynamic characterization of
the ocular vasculature.
1,2
This technique is a non-
invasive, reproducible, and easily applied method of
examining hemodynamic changes in several orbital
and retinal vascular diseases.
35
The earliest detectable
morphologic changes in diabetic retinopathy are
microaneurysms and capillary closure. Proliferation
of vascular endothelial cells in proliferative diabetic
retinopathy represents ischemia of the inner retinal
layers secondary to closure of portions of the retinal
capillary bed.
6
However, the pathogenic mechanisms
important to the initiation and progression of diabetic
retinopathy are not fully understood. Using color
Doppler sonography, Mendivil and associates
7
and
Mendivil and Cuartero
8
reported that ocular blood
flow velocity was decreased in diabetic patients with
proliferative diabetic retinopathy as compared to nor-
mal subjects and was significantly decreased after
scatter photocoagulation. Goebel and coworkers
9
reported a correlation between the severity of diabetic
retinopathy and decreased flow velocity in the central
Received April 6, 1998, from The Third Department of Internal
Medicine, Yokohama City University School of Medicine, Yokohama,
Japan. Revised manuscript accepted for publication August 2, 1998.
Address correspondence and reprint requests to Kazushi Numata,
MD, The Third Department of Internal Medicine, Yokohama City
University School of Medicine, 3-9 Fukuura, Kanazawa-ku,
Yokohama 236, Japan.
Acknowledgment: Financially supported in part by a grant
(09770370) from the Ministry of Education, Science, Sports, and
Culture of Japan.
ABBREVIATIONS
RI, Resistive index; SD, Standard deviation; PSV, Peak systolic
velocity; EDV, End diastolic velocity; ANOVA, Analysis of vari-
ance
1998 by the American Institute of Ultrasound in Medicine J Ultrasound Med 17:675681, 1998 0278-4297/98/$3.50
R
Ocular Arterial Flow Hemodynamics in
Patients with Diabetes Mellitus
Takashi Arai, MD, Kazushi Numata, MD, Katsuaki Tanaka, MD, Takayoshi Kiba, MD,
Satsuki Kawasaki, MD, Tatsuya Saito, MD, Shinobu Satoh, MD, Hisahiko Sekihara, MD
The purpose of this study was to investigate ocular
blood flow hemodynamics in patients with diabetes
mellitus. We used color Doppler sonography, in 22
normal subjects and 52 patients with (n = 25) or with-
out (n = 27) diabetic retinopathy, to determine blood
flow velocities and the resistive index of the central
retinal artery. The resistive index of the central retinal
artery in patients with diabetic retinopathy (0.85
0.09) was significantly greater (P < 0.01) than that in
normal subjects (0.72 0.08) and in patients without
diabetic retinopathy (0.81 0.09). The resistive index
of the central retinal artery in the patients without
diabetic retinopathy was also significantly greater
than that of normal subjects (P < 0.01). The resistive
index of ocular arterial flow was increased in the
patients with diabetes mellitus and further increased
in the presence of retinopathy. Increased resistance in
the peripheral ocular vascular bed contributes to dia-
betic retinopathy, and this change is present before
the appearance of overt diabetic retinopathy. KEY
WORDS: Color Doppler sonography; Central retinal
artery; Diabetic retinopathy; Resistive index.
retinal artery. Tamaki and colleagues
10
measured
ophthalmic arterial flow velocities in diabetic and
normal eyes and found that, in the diabetic group,
the RI was significantly higher than that in the con-
trol group and increased with the severity of
retinopathy. Elevation in the RI of the ophthalmic
artery may be due to increased resistance in the dis-
tal vascular bed. However, the RI of the central reti-
nal artery in diabetic patients has not been
investigated previously.
Using color Doppler sonography to assess the
orbital hemodynamics, we investigated the velocities
and the RIs of both the ophthalmic and the central
retinal artery in normal subjects and patients with or
without diabetic retinopathy. The purpose of our
study was to evaluate resistive changes in the
peripheral ocular vascular bed that may contribute
to the initiation of diabetic retinopathy and to assess
the usefulness of color Doppler sonography for dif-
ferentiating patients with diabetic retinopathy from
those without diabetic retinopathy.
SUBJECTS AND METHODS
Subjects
The study population consisted of 55 diabetic patients
and 22 sex- and age-matched control subjects.
Diabetic retinopathy was classified according to the
modified Airlie House system
11,12
on the basis of
stereoscopic fundus examination with a 90 diopter
lens. We excluded three diabetic patients with poorly
controlled systemic hypertension from this study.
Therefore, we studied 54 eyes of 27 patients without
diabetic retinopathy (mean age SD: 50.0 17.3 years;
15 men, 12 women), 50 eyes of 25 patients with dia-
betic retinopathy (eight patients with background
retinopathy, three patients with proliferative retinopa-
thy, and 14 patients with postpanretinal photocoagu-
lation) (mean age SD: 56.1 9.9 years; 14 men, 11
women), and 44 eyes of 22 normal subjects (mean age
SD: 49.9 18.9 years; 12 men, 10 women). All the
diabetic patients had noninsulin-dependent diabetes
mellitus. No significant difference was found in the
hemoglobin A
1c
level between the patients without
diabetic retinopathy (mean SD: 8.8 2.8%) and
those with diabetic retinopathy (mean SD: 8.3
2.1%). None of the normal subjects was taking med-
ications or had a previous history of diabetes mellitus,
and all had normal glucose function tests. Informed
consent was obtained from all participants, and the
study protocol was approved by the institutional
ethics committee. The sonographers had no informa-
tion on the patients at the time of the examinations.
Equipment
All Doppler sonographic examinations were per-
formed with a Sonoline Elegra and Quantum 2000
color Doppler sonography unit (Siemens Medical
Systems, Issaquah, WA) using a 7.5 MHz linear probe.
Procedures
Ultrasonographic scanning was performed with
the patient in the supine position with closed eyes.
Color Doppler sonography was used to detect the
ophthalmic and central retinal arteries, and a
Doppler waveform was examined in this region
while the measurements were made (Fig. 1Aand B).
The color Doppler display was set so that blood flow
toward the transducer appeared in red, while blood
flow away from the transducer appeared in blue. The
threshold levels were set to optimize sensitivity with-
out producing excessive system noise. Examinations
were performed at a variety of flow settings, depend-
ing upon the underlying flow velocities.
Real-time color images of the ophthalmic and cen-
tral retinal arteries were obtained, followed by the
Doppler spectral analysis and velocity calculations
of flow in the arteries. The PSV and EDV in the oph-
thalmic and central retinal arteries were determined
after correction for the angle of insonation. The angle
was always less than 60 degrees, and every measure-
ment was made at a constant angle to the vessel. PSV
represents the maximum flow velocity recorded dur-
ing each cardiac cycle. EDV was recorded immedi-
ately before the next systolic upstroke.
13
The RIs
[(PSV EDV)/PSV] of the ophthalmic and central
retinal arteries were calculated. Because the oph-
thalmic and central retinal arteries were too small in
diameter to permit precise measurement, the cross-
sectional areas of these vessels were not calculated.
To reduce variability, the results represent the aver-
age of at least three measurements.
14
The intraindi-
vidual coefficient of variation for every sonographic
parameter was less than 10%. Blood pressure was
monitored throughout the study, and all subjects had
systolic pressures below 140 mm Hg and diastolic
blood pressures below 90 mm Hg.
Statistical Analysis
Data are expressed as the mean SD. Results were
analyzed statistically by ANOVA. When the F values
proved significant, Fishers multiple comparison test
was used to identify significant differences between
the groups. Differences within groups were evalu-
ated by the paired t-test. AP value less than 0.05 was
considered statistically significant.
676 OCULAR ARTERIAL FLOW HEMODYNAMICS J Ultrasound Med 17:675681, 1998
RESULTS
Ophthalmic Artery
Table 1 presents the mean values of the flow veloci-
ties and the RIs in normal subjects and diabetic
patients. EDVs in the diabetic patients, regardless of
the presence or absence of retinopathy, were signifi-
cantly lower than those of normal subjects (P < 0.05
and 0.01, respectively). EDVs in the patients with
diabetic retinopathy were significantly lower than
those of patients without retinopathy (P < 0.05).
However, no differences were seen in PSVs among
the three groups. The RIs in patients with diabetic
retinopathy (0.86 0.06) were significantly greater
(P < 0.01) than those in normal subjects (0.78 0.07)
and in patients without diabetic retinopathy (0.81
0.06). No significant difference in the RI was noted
between patients without diabetic retinopathy and
normal subjects. When the RI was 0.80 or greater,
sensitivity was 72%, specificity 57%, and accuracy
66% in distinguishing normal subjects from diabetic
patients. When the RI was 0.86 or greater, sensitivity
was 56%, specificity 72%, and accuracy 64% in dis-
tinguishing patients with diabetic retinopathy from
those without diabetic retinopathy.
Central Retinal Artery
Table 2 presents the mean values of the flow veloci-
ties and the RIs in normal subjects and diabetic
patients. The PSVs and EDVS in the diabetic
patients, regardless of retinopathy, were signifi-
cantly lower than those of normal subjects (P < 0.01).
A significant difference in EDV was noted between
the patients with and the patients without diabetic
677 J Ultrasound Med 17:675681, 1998 ARAI ET AL
Figure 1 Representative orbital artery sonograms from a normal subject and a diabetic patient. A, Color Doppler sonographic
image shows the ophthalmic artery (arrowhead). B, Color Doppler sonographic image shows the central retinal artery (arrow).
A B
Table 1: Velocities (cm/s) and the RIs of the Ophthalmic Artery in Normal Subjects and in Patients with
Diabetes
Patients with Diabetes
Parameters Normal Subjects Retinopathy () Retinopathy (+) Total Patients
No. of subjects (eyes) n = 44 n = 54 n = 50 n = 104
PSV 32.1 11.3 30.6 9.7 27.2 11.7 29.0 10.7
EDV 7.2 4.2 5.7 2.5* 3.8 2.5 4.8 2.7
RI 0.78 0.07 0.81 0.06 0.86 0.06 0.84 0.07
Values are expressed as the mean SD.
*P < 0.05.
P < 0.01 compared with normal subjects.
P < 0.05.
P < 0.01 compared with the patients without diabetic retinopathy.
retinopathy (P < 0.01). No difference in PSV was
found between the patients with and those without
diabetic retinopathy. The RI in the patients without
diabetic retinopathy was significantly greater than
that of normal subjects (P < 0.01). The RI in the
patients with diabetic retinopathy (0.85 0.09) was
significantly greater (P < 0.01) than that of normal
subjects (0.72 0.08) and of patients without diabetic
retinopathy (0.81 0.09) (Fig. 2Aand B). When the RI
was 0.80 or greater, sensitivity was 66%, specificity
88%, and accuracy 70% in distinguishing normal
subjects from diabetic patients. When the RI was 0.85
or greater, sensitivity was 64%, specificity 63%, and
accuracy 63% in distinguishing patients with dia-
betic retinopathy from those without diabetic
retinopathy.
DISCUSSION
Color Doppler sonography is useful for investigating
the hemodynamics (i.e., flow) in small arteries.
14,1517
In examining ocular blood flow, this tool can be used
when standard diagnostic procedures (e.g., fundus
examination or fluorescein angiography) are difficult
because of cataract and vitreous hemorrhage.
9
Using
this technology, we showed the RIs of ocular arterial
flow in patients with diabetes mellitus to be signifi-
cantly greater than those of normal subjects and to be
further increased in the presence of diabetic
retinopathy. Only in the central retinal artery was the
RI in patients without diabetic retinopathy signifi-
cantly greater than that of normal subjects. The PSVs
and EDVs of the central retinal artery in the diabetic
patients, regardless of the presence or absence of
retinopathy, were significantly lower than those of
normal subjects. A significant difference in the EDV
of the central retinal artery was noted between the
patients with and those without diabetic retinopathy.
However, no differences in the PSVs of the oph-
thalmic artery were noted among the three groups.
EDVs of the ophthalmic artery in the patients with
diabetic retinopathy were significantly lower than
those of patients without retinopathy. Therefore,
measurements of these parameters in the central reti-
nal artery are superior to those of the ophthalmic
artery in terms of assessing the ocular hemodynamic
changes in diabetes and for detecting the presence of
diabetic retinopathy. Because the central retinal
artery constitutes a peripheral portion of the ocular
vasculature as compared to the ophthalmic artery,
this artery reflects more sensitively the vascular
changes associated with diabetes.
In previous studies, the severity of diabetic
retinopathy was examined by measuring the blood
flow velocities in the ophthalmic artery and the cen-
tral retinal artery using color Doppler sonogra-
phy.
79,18
In this study, no significant differences in the
PSV of the ophthalmic artery were found among
normal subjects and patients with or without dia-
betic retinopathy, while the RIs in patients with dia-
betic retinopathy were significantly greater than
those of normal subjects and patients without dia-
betic retinopathy. Therefore, we consider measure-
ments of the RI in the ophthalmic and central retinal
arteries to be better parameters than the velocities of
the blood flow in these vessels for evaluating the
severity of diabetic retinopathy.
The RI has been used as a measure of vascular
resistance in the artery.
1921
Tamaki and colleagues
10
measured the RI of the ophthalmic artery in diabetic
and normal eyes and found that, in the diabetic
group, the RI was significantly higher than in the
control group and increased with the severity of
retinopathy. Guven and associates
18
demonstrated
the absence of a diastolic flow pattern or an
extremely low diastolic velocity of the central retinal
artery in nine of 73 eyes and suggested these findings
678 OCULAR ARTERIAL FLOW HEMODYNAMICS J Ultrasound Med 17:675681, 1998
Table 2: Velocities (cm/s) and the RIs of the Ophthalmic Artery in Normal Subjects and in Patients with
Diabetes
Patients with Diabetes
Parameters Normal Subjects Retinopathy () Retinopathy (+) Total Patients
No. of subjects (eyes) n = 44 n = 54 n = 50 n = 104
PSV 12.3 4.0 9.4 2.6* 8.6 3.5* 9.0 3.1*
EDV 3.5 1.6 1.9 1.2* 1.3 1.1* 1.6 1.2*
RI 0.72 0.08 0.81 0.09* 0.86 0.06* 0.84 0.07*
Values are expressed as the mean SD.
*P < 0.01 compared with normal subjects.
P < 0.01 compared with the patients without diabetic retinopathy.
to be due to increased resistance in the distal vascu-
lar bed. Langham and coworkers,
22
using a noninva-
sive computerized methodology, showed choroidal
blood flow decreases to parallel the increasing sever-
ity of retinopathy in diabetic patients because of
increased vascular resistance. Previous studies found
that in the early stage of diabetes some patients
showed histopathologic changes (e.g., thickening of
the basement membranes of choroidal capillaries,
narrowing or obstructive changes in the choroidal
capillaries, loss of pericytes, and increased suben-
dothelial hyaline deposition) all of which contribute
to increased resistance in the peripheral vascular
bed.
2325
Garner and Ashton
26
reported that decreases
in the internal space of the ophthalmic artery also
contributed to rising vascular resistance in patients
with diabetic retinopathy. Blood rheologic abnormal-
ities tend to compromise the antithrombotic and fib-
rinolytic activity of the vascular endothelium,
leading to the obstruction of retinal capillaries.
27
In our study, the RIs of the ophthalmic and the
central retinal arteries in patients without diabetic
retinopathy were significantly greater than those of
normal subjects. Other investigators have also
reported the RI of the ophthalmic artery in patients
without diabetic retinopathy to be significantly
higher than that of normal subjects.
8
These results
suggest that increased resistance in the peripheral
vascular bed might be initiated prior to the appear-
ance of overt diabetic retinopathy. Several studies
have suggested this possibility.
28,29
Evans and associ-
ates
28
reported that normal subjects experienced
reduced ophthalmic and central retinal arterial
blood flow velocity during hyperoxia, while insulin-
dependent diabetic patients showed no change.
Their results demonstrate that diabetic patients with
minimal or no retinopathy show ocular vascular
function irregularities in the major vessels feeding the
eye. Using a laser Doppler technique, Feke and
coworkers
29
also reported that retinal arterial blood
679 J Ultrasound Med 17:675681, 1998 ARAI ET AL
Figure 2 Representative central retinal artery sonograms from a normal subject and a diabetic patient. A, Color Doppler sono-
graphic image shows the central retinal artery in a normal subject. B, Color Doppler sonographic image shows the central
retinal artery in a patient with diabetic retinopathy. Spectral analysis shows a markedly reduced end diastolic flow velocity
as compared to that seen in A.
A B
velocities in insulin-dependent diabetic patients were
already low before the clinical appearance of
retinopathy.
In this study, we did not measure blood flow vol-
ume in the ophthalmic and central retinal arteries,
because these vessels were too small in diameter to
permit precise measurement of the cross-sectional
area. Laser Doppler velocimetry can be used to mea-
sure velocity, cross-sectional area, and flow volume in
the retinal artery. Using this technology, retinal arter-
ial flow velocity and volume in patients without dia-
betic retinopathy, with background retinopathy,
29
and
with postpanretinal photocoagulation
30
were shown
to be decreased as compared to those in normal sub-
jects. Laser photocoagulation treatment reduced reti-
nal blood flow velocity and volume.
31
In contrast,
retinal arterial cross-sectional area and flow volume
in patients with proliferative retinopathy were
increased as compared to those in normal subjects.
30,32
These results are consistent with those of our study.
Among our patients with diabetic retinopathy, those
with proliferative retinopathy were few, and most of
our patients had background retinopathy or postpan-
retinal photocoagulation. Blood flow velocity
through the central retinal artery was significantly
lower in those with diabetic retinopathy than in those
without diabetic retinopathy and in normal subjects.
Our study had several limitations. First, few of our
patients with diabetic retinopathy had proliferative
retinopathy; most had background retinopathy or
postpanretinal photocoagulation. If the postpanreti-
nal photocoagulation patients were not included in
this study, the value of the RI of the central retinal
artery in the patients with diabetic retinopathy might
decrease, because destruction of small vesssels by
panretinal photocoagulation probably caused an
increase in vascular resistance in retinal blood flow
and contributed to the increase in the RI.
33
Second,
our study population was relatively small. There-
fore, the accuracy of the RI for differentiating
between patients with and without diabetic
retinopathy was relatively low. A large series might
have disclosed additional features related to the use-
fulness of color Doppler sonography for distinguish-
ing patients with diabetic retinopathy from those
without diabetic retinopathy.
Although further study of orbital blood flow in dia-
betic retinopathy is needed, our results suggest that
color Doppler sonography has the potential to pro-
vide useful information about hemodynamic changes
in patients with diabetic retinopathy. Diabetic
retinopathy may be triggered by increased resistances
in the peripheral ocular vascular bed, changes which
may begin in the early stage of diabetes.
REFERENCES
1. Guthoff RF, Berger RW, Winkler P, et al: Doppler sonog-
raphy of the ophthalmic and central retinal vessels. Arch
Ophthalmol 109:532, 1991
2. Lieb WE, Cohen SM, Merton DA, et al: Color Doppler
imaging of the eye and orbit. Arch Ophthalmol 109:527,
1991
3. Erickson SJ, Hendrix LE, Massaro BM, et al: Color
Doppler flow imaging of the normal and abnormal orbit.
Radiology 173:511, 1989
4. Lieb WE: Color Doppler ultrasonography of the eye and
orbit. Curr Opinion Ophthalmol 4:68, 1993
5. Regillo CD, Sergott RC, Ho AC, et al: Hemodynamic
alterations in the acute retinal necrosis syndrome.
Ophthalmology 100:1171, 1993
6. Glaser B: Extracellular modulating factors and the control
of intraocular neovascularization. Arch Ophthalmol
106:603, 1988
7. Mendivil A, Cuartero V, Mendivil MP: Ocular blood flow
velocities in patients with proliferative diabetic retinopa-
thy and healthy volunteers: A prospective study. Br J
Ophthalmol 79:413, 1995
8. Mendivil A, Cuartero V: Ocular blood flow velocities in
patients with proliferative diabetic retinopathy after scat-
ter photocoagulation: Two years of follow-up. Retina
16:222, 1996
9. Goebel W, Lieb WE, Ho A, et al: Color Doppler imaging:
A new technique to assess orbital blood flow in patients
with diabetic retinopathy. Invest Ophthalmol Vis Sci
36:864, 1995
10. Tamaki Y, Nagahara M, Yamashita H, et al: Blood veloci-
ty in the ophthalmic artery determined by color Doppler
imaging in normal subjects and diabetes. Jpn J
Ophthalmol 37:385, 1993
11. The Diabetic Retinopathy Study Research Group: Amod-
ification of the Airlie House classification of diabetic
retinopathy: DRS Report #7. Invest Ophthalmol Vis Sci
21:210, 1981
12. Early Treatment Diabetic Retinopathy Study Research
Group: Grading diabetic retinopathy from stereoscopic
color fundus photographs: An extension of the modified
Airlie House classification: ETDRS Report #10.
Ophthalmology 98:786, 1991
13. Moneta GL, Taylor DC, Helton WS, et al: Duplex ultra-
sound measurement of postprandial intestinal blood
flow: Effect of meal composition. Gastroenterology
95:1294, 1988
14. Paulson EK, Kliewer MA, Frederick MG, et al: Hepatic
artery: Variability in measurement of resistive index and
systolic acceleration time in healthy volunteers.
Radiology 200:725, 1996
15. Numata K, Tanaka K, Kiba T, et al: Use of hepatic tumor
index on color Doppler sonography for differentiating
large hepatic tumors. AJR 168:991, 1997
16. Tanaka K, Mitui K, Morimoto M, et al: Increased hepatic
arterial blood flow in acute viral assessment by color
Doppler sonography. Hepatology 18:21, 1993
680 OCULAR ARTERIAL FLOW HEMODYNAMICS J Ultrasound Med 17:675681, 1998
17. Numata K, Tanaka K, Mitsui K, et al: Flow characteristics
of hepatic tumors at color Doppler sonography:
Correlation with arteriographic findings. AJR 160:515,
1993
18. Guven D, Ozdemir H, Hasanreisoglu B: Hemodynamic
alterations in diabetic retinopathy. Ophthalmology
103:1245, 1996
19. Sacerdoti D, Merkel C, Bolognesi M, et al: Hepatic arteri-
al resistance in cirrhosis with and without portal vein
thrombosis: Relationships with portal hemodynamics.
Gastroenterology 108:1152, 1995
20. Norris CS, Barnes RW: Renal artery flow velocity analy-
sis: Asensitive measure of experimental and clinical ren-
ovascular resistance. J Surg Res 36:230, 1984
21. Gosling RG, Dunbar G, King DH, et al: The quantitative
analysis of occlusive arterial disease by a nonintrusive
ultrasonic technique. Angiology 22:52, 1971
22. Langham ME, Grebe R, Hopkins S, et al: Choroidal blood
flow in diabetic retinopathy. Exp Eye Res 52:167, 1991
23. Yanoff M: Ocular pathology of diabetes mellitus. Am J
Ophthalmol 67:21, 1969
24. Hidayat AA, Fine BS: Diabetic choroidopathy: Light and
electron microscopic observation of seven cases.
Ophthalmology 92:512, 1985
25. Takebayashi S, Ogata J, Jimi S: Clinical choroidal throm-
bosis, hypertension, and diabetes mellitus: An electron
microscopic study. Hum Pathol 19:99, 1988
26. Garner A, Ashton N: Ophthalmic artery stenosis and dia-
betic retinopathy. Trans Ophthalmol Soc UK 7:101, 1972
27. Kohner EM, Porta M: Vascular abnormalities in diabetes
and their treatment. Trans Ophthalmol Soc UK 100:440,
1980
28. Evans DW, Harris A, Danis RP, et al: Altered retrobulbar
vascular reactivity in early diabetic retinopathy. Br J
Ophthalmol 81:279, 1997
29. Feke GT, Buzney SM, Ogasawara H, et al: Retinal circula-
tory abnormalities in type I diabetes mellitus. Invest
Ophthalmol Vis Sci 35:2968, 1994
30. Patel V, Rassam S, Newsom R, et al: Retinal blood flow in
diabetic retinopathy. Br Med J 305:678, 1992
31. Fujio N, Feke GT, Gorger DG, et al: Regional retinal blood
flow reduction following half-fundus photocoagulation
treatment. Br J Ophthalmol 78:335, 1994
32. Konno S, Feke GT, Yoshida A, et al: Retinal blood flow
changes in type I diabetes. Along-term follow-up study.
Invest Ophthalmol Vis Sci 37:1140, 1996
33. Grunwald JE, Brucker AJ, Petrig BL, et al: Retinal blood
flow regulation and the clinical response to panretinal
photocoagulation in proliferative diabetic retinopathy.
Ophthalmology 96:1518, 1989
681 J Ultrasound Med 17:675681, 1998 ARAI ET AL