Colour Doppler imaging of the ocular circulation ... - Wiley Online Library

A O S 2000

Colour Doppler imaging of the ocular circulation in diabetic retinopathy Jane R. MacKinnon1, Graham McKillop2, Colm O’Brien1, Kenneth Swa1, Zahida Butt1 and Patricia Nelson1 1

Princess Alexandra Eye Pavilion and 2Radiology Department, Royal Infirmary of Edinburgh NHS Trust, Scotland

ABSTRACT. Purpose: To measure blood flow velocity in the ophthalmic artery (OA) and central retinal artery (CRA) in patients with diabetic retinopathy. Subjects and Methods: 62 age-matched subjects divided into 3 groups: nondiabetic controls (nΩ17); diabetics with no clinical retinopathy or background changes (nΩ24); diabetics with either pre-proliferative or proliferative retinopathy (nΩ21). Colour Doppler imaging was performed on supine patients by one masked observer using the Acuson 128 machine. Results: There was a statistically significant (p∞0.05) decrease in both the peak systolic velocity (PSV 0.073 m/s) and end diastolic velocity (EDV 0.014 m/s) of the central retinal artery in the pre-proliferative/proliferative group compared to the no retinopathy/background retinopathy group (PSV 0.096 m/s, EDV 0.024 m/s) and the control group (PSV 0.142 m/s, EDV 0.029 m/s). The resistance index of the ophthalmic artery was significantly increased in both the preproliferative/proliferative (0.81) and the no retinopathy/background group (0.81) compared to controls (0.72). Conclusion: Reduced blood flow velocity was found in the CRA of diabetic patients and appeared to become further reduced with the progression of retinopathy. This suggests that monitoring with Colour Doppler imaging may have predictive power in identifying those at greatest risk of developing sight threatening proliferative disease. The resistance index of the OA was increased in diabetics compared to controls.

have been relatively inaccessible to study in vivo. The introduction of orbital colour Doppler imaging (CDI) in 1989 (Erickson et al. 1989) presented the opportunity for assessment of orbital blood vessels. CDI is a non-invasive ultrasonic method for qualitatively and quantitatively assessing blood flow velocity information (Aburn & Sergott 1993; Lieb et al. 1991). Good reproducibility of measurements in the ophthalmic and central retinal arteries has been reported (Baxter & Williamson 1995). Simultaneous Doppler information is gathered from a cross-section of tissue combined with Bscan recording. Pulsatile blood velocity profiles are then obtained and analysed. The purpose of this study was to measure blood flow velocity in the ophthalmic and central retinal artery in patients with untreated diabetic retinopathy and compare the results with age matched normal control subjects.

Key words: Colour Doppler imaging – diabetic retinopathy – ocular circulation – central retinal artery – ophthalmic artery.

Subjects and Methods

Acta Ophthalmol. Scand. 2000: 78: 386–389 Copyright c Acta Ophthalmol Scand 2000. ISSN 1395-3907

D

iabetes mellitus is a condition of altered metabolism resulting in systemic vascular disease. The pathogenesis of diabetes and in particular its effect on the ocular circulation is not fully understood. Diabetic retinopathy is a major cause of blindness in the United Kingdom, resulting in approximately 1000 blind registrations per year (Ulbig & Hamilton 1993). Diabetic retinopathy is a microangiopathy which ultimately may result in retinal ischaemia (Patel et al. 1992). This is thought to be the stimulus

386

for neovascularisation and therefore the visually threatening events of vitreous haemorrhage, glaucoma and tractional retinal detachment (Forrester et al. 1993; Kohner 1993). The vascular effects of diabetes on the retinal, and to a lesser extent, the choroidal circulation have been studied with conflicting evidence as to whether blood flow is increased (Arend et al. 1991; Patel et al. 1992) or decreased (Sinclair 1991; Feke et al. 1994). The characteristics of the vasculature proximal to these vessels

Seventeen non-diabetic subjects and 45 diabetic patients were included in the study. The non-diabetic controls were recruited from hospital staff and voluntary workers. Diabetes was diagnosed by at least 2 fasting plasma glucose measurements above 7.8 mmol/l; patients were grouped into Type 1 and 2 on the basis of their clinical presentation and features (Kumar & Clark 1998). The diabetic subjects were recruited prospectively from the diabetic retinopathy clinic and were divided into two separate sub-groups: 1) no diabetic retinopathy (NDR, nΩ12) and background retinopathy (BDR, nΩ 12); 2) pre-proliferative retinopathy

A O S 2000 (PPDR, nΩ15) or proliferative diabetic retinopathy (PDR, nΩ6). The latter two groups were pooled because of the small patient numbers in the proliferative group. The diabetic subjects were categorised following dilated fundoscopy using the biomicroscopic indirect ophthalmoscope by a single experienced medical ophthalmologist (KS). Patients were classified according to the modified Airlie House system (DRS report 1981 and ETDRS report 1991). Patients with NDR had no clinically observable diabetic retinopathy. BDR was defined by the presence of microaneurysms, haemorrhages and exudates (Airlie House retinopathy classification 2–3). Pre-proliferative changes consisted of background retinopathy lesions plus two or more of the following: venous beading or reduplication, intra-retinal microvascular abnormalities, deep intra-retinal haemorrhages and multiple cotton wool spots (corresponding to retinopathy level 4–5). Proliferative diabetic retinopathy was characterised by areas of papillary and/or epiretinal new vessel formation (retinopathy level 6–7). In the NDR/BDR group there were 7 Type 1 and 17 Type 2 diabetics. There was a similar proportion of diabetic types in the PPDR/PDR group – 6 Type 1 and 15 Type 2 (p±0.05 using chi squared analysis). There was no significant difference in the duration of diabetes between the 2 diabetic groups (ANOVA). The time from formal diagnosis varied from 2 months to 32 years in the NDR/BDR group, the average being 10 years; the average duration was 12 years in the PPDR/ PDR group, range 3 months to 34 years. Excluded from the study were patients who had undergone previous laser photocoagulation. Our exclusion criteria also prevented those with a history of any eye disease which may affect blood flow being studied, such as ocular inflammation, trauma or abnormality, non-diabetic vascular disease (retinal vein occlusion, for example), or glaucoma. Information on medication use was available from all subjects. None of the controls were on any cardiac or anti-hypertensive drugs while 29 out of the 45 diabetics (64%) were on single or multiple agents. Comparing the 2 diabetic groups, significantly more (chi test 0.006) of those with advanced retinopathy were on systemic antihypertensives (16/21, 76%) than in the lesser affected group (13/24, 54%). However, there were no significant differences in the proportions prescibed b-

blockers, calcium channel antagonists or ACE inhibitors between the groups (chi test ±0.3). Colour Doppler imaging of the ophthalmic artery (OA) and the central retinal artery (CRA) was performed by a single masked experienced observer (G McK) using the Acuson 128 machine (Mountain View, CA). After the procedure had been explained to the subject they assumed a supine position and a 7.5 MHz probe was applied to the closed eyelids using sterile coupling gel. The examination technique and intra-individual variation have been described previously (Butt et al. 1995). Angle correction was applied to the pulsed Doppler recordings to minimize errors in the measured velocities. Measurements of peak systolic velocity (PSV, m/sec) and end diastolic velocity (EDV, m/sec) were obtained using the mean of 3 cardiac cycles. Results from a single eye were used, chosen randomly by tossing a coin, if both eyes were equally involved. In the PDR group the eye with active neovascularisation was chosen. The resistance index (RI) was then calculated according to Pourcelot’s formula (Planiol et al. 1972): RIΩ

PSVªEDV PSV

Stastistical analysis A one way analysis of variance (ANOVA) across the 3 groups was carried out for each set of variables followed where appropriate with t-tests using the Bonferroni correction (Winer et al. 1991). Data which was found not to be normally distributed was analysed using the KruskalWallis ANOVA and Mann-Whitney test (Norusis 1993). Correlation between variables was looked for using the Spearman coefficient (Norusis 1993). Chi squared analysis was used where indicated. Statistical significance was set at p∞0.05.

Results Table 1 summarizes the clinical data. There was no significant difference in the age of the subjects (ANOVA). Chi squared analysis showed a significant difference in the sex ratio between the 3 groups (p⬍0.05) with the control group having a higher proportion of females and the PPDR/PDR group having more males. Table 2 summarizes the CDI results. The ANOVA indicated significant differences in the resistance index of the OA (FΩ14.73, dfΩ2, pΩ0.0001). Follow-up ttests with Bonferroni correction showed

Table 1. Systemic data. Mean values shown, standard deviation in brackets.

Number Age (years) Sex (male/female)

Controls

NDR/BDR

PPDR/PDR

17 66 (5) 4/13

24 62 (13) 14/10

21 61 (9) 16/5

Table 2. Colour Doppler imaging results for the ophthalmic artery and central retinal artery. Mean values shown, SD in brackets. * Denotes significant value compared to control and † denotes significant value compared to NDR/BDR group (see text for details). Controls (nΩ17)

NDR/BDR (nΩ24)

PPDR/PDR (nΩ21)

Ophthalmic artery PSV (m/s) EDV (m/s) RI

0.31 (0.10) 0.09 (0.03) 0.72 (0.05)

0.36 (0.13) 0.07 (0.04) 0.81 (0.06)*

0.34 (0.13) 0.07 (0.04) 0.81 (0.07)*

Central retinal artery PSV (m/s) EDV (m/s) RI

0.14 (0.07) 0.03 (0.02) 0.80 (0.09)

0.10 (0.04)* 0.02 (0.02) 0.78 (0.06)

0.07 (0.03)*† 0.01 (0.01)*† 0.82 (0.06)

Abbreviations: NDR – no diabetic retinopathy; BDR – background diabetic retinopathy; PPDR – pre-proliferative diabetic retinopathy; PDR – proliferative diabetic retinopathy; PSV – peak systolic velocity; EDV – end diatolic velocity; RI – resistance index.

387

A O S 2000 a significantly higher resistance index in the OA of both the NDR/BDR and the PPDR/PDR groups (pΩ0.0001) compared to the normal subjects. The Kruskal-Wallis ANOVA indicated significant differences in the PSV (dfΩ2, pΩ0.0002) and EDV (dfΩ2, pΩ0.003) of the CRA. Follow-up Mann – Whitney tests indicated significantly lower PSV in the CRA of both the NDR/BDR group (pΩ0.02) and the PPDR/PDR group (pΩ0.0001) compared with the normal subjects. There was also a significantly lower EDV in the CRA of the PPDR/PDR group (pΩ0.005) compared with the control group. In the 2 diabetic groups there was a significantly lower central retinal artery PSV (pΩ0.02) and EDV (pΩ0.004) in the PPDR/PDR group compared with the NDR/BDR group. The Mann-Whitney test was used to compare the CDI values for those subjects with PPDR against those with PDR: no significant difference was found between the two groups.

Discussion Compared with the normal control subjects, diabetic patients were found to have a significantly higher resistance index in the OA, and slower velocities in the CRA. Those with PPDR/PDR also had significantly lower velocities in the CRA compared to diabetics with NDR/BDR, ie., the more severe the disease category the slower the arterial velocity in the CRA. Our results agree with the study from Goebel et al. (1995), which found significantly lower PSV and EDV in the CRA of all those with untreated diabetic retinopathy compared to controls; those with PDR also had a significantly reduced PSV compared to BDR subjects. They were unable to show a significant difference in the velocity of the OA and short posterior ciliary arteries. RI was not measured in their study. Similar results were found by Mendivil et al. (1995), who compared a group of 25 diabetics (type 1 and type 2) with PDR with 30 non-diabetic controls. As well as finding lower systolic, diastolic and mean velocities in the CRA they also report that these parameters were significantly reduced in the OA. CDI values in diabetics with non-proliferative retinopathy were not reported in Mendivil’s paper. These results differ from those in our study in that we found no significant difference in the OA velocities. Our OA data

388

tends to indicate an increase in PSV and a decrease in EDV in the diabetic groups compared to controls. A decrease in EDV suggests downstream impedence (Rankin et al. 1995), while an increase in PSV may be associated with increases in systolic blood pressure (Williamson et al. 1995). It has previously been shown by Williamson that SBP is positively correlated with PSV in the OA and CRA in normals. Our results do not confirm this correlation, either collectively or in group analysis. There were no significant differences between the SBP’s of our subjects (ANOVA), although this should be interpreted with caution as blood pressure measurements were only obtained once (Armitage & Rose 1966). Resistance index was not measured in either of the above studies (Goebel et al. 1995; Mendivil et al. 1995) of CDI in diabetic retinopathy. The value of calculating the resistance index has previously been highlighted by Mulhern et al. (1996), helping to give a more accurate and non-angle dependent measurement. Our finding of increased RI in the OA has not been previously reported to our knowledge. The increase in RI is due to a statistically nonsignificant increase in PSV and a decrease in EDV. It is suggested that an increased RI in the OA in both diabetic groups may be due to downstream vascular changes, e.g., diabetes associated vascular damage in both the retinal and choroidal vasculature. Measurement of OA blood flow velocity probably represents changes in choroidal blood flow. It has been estimated that of the total ocular blood supply less than 10% flows to the retina (Hart 1992), the remainder being directed to the choroid via the ciliary circulation. McLeod and Lutty (1994) have demonstrated significant angiopathic changes in the diabetic choroid, including extensive capillary dropout, beaded capillaries and neovascularisation. We have previously reported that pulsatile ocular blood flow (POBF), which is believed to be an indirect measure of choroidal blood flow, is increased in diabetics (MacKinnon et al. 1997). It is suggested that the finding of a wider pulse velocity in the OA of diabetics may correlate with an increased POBF. Because diabetes mellitus is a systemic vascular disease many of our subjects (29/ 45, 64%) were on medication with vascular effects, whereas none of the controls were. More of those in the PPDR/PDR group took antihypertensive drugs than those in the NDR/BDR group (76% versus 54%). This is of importance in interpretation of

our results as many of the agents prescribed will cause a change in vascular tone. A change in vessel diameter affects the blood velocity according to Poiseulles law. The finding of reduced velocities in the CRA may therefore be related to drug effect. Our study would have been enhanced by matching patients with controls on the same medication. CDI has also given some interesting information on the changes in ocular blood flow velocities following retinal laser treatment. Mendivil & Cuartero (1996) have reported on the CDI findings of a mixed group of type 1 and type 2 diabetics with PDR after scatter laser photocoagulation. Panretinal photocoagulation resulted in a reduction in the PSV of the OA compared to values before laser treatment. These results were obtained 6 months following treatment and did not change during 2 years of follow-up. Evans et al. (1997), also used CDI to investigate retrobulbar vascular reactivity in early diabetic retinopathy. CDI of the OA and CRA was performed on 11 patients with early or no diabetic retinopathy and compared to 11 non-diabetic controls. While under conditions of isocapnic hyperoxia, diabetics exhibited a significantly lower resistance index in both the OA and CRA compared to controls. Hypoxia also induced a higher EDV in the CRA and lower PSV in the OA compared to normals. Grunwald et al. (1996) evaluated fundus photographs and confirmed a statistically significant enlargement of venous and arterial vessel diameter in patients with insulin dependent diabetes compared to normal controls. Vasodilation results in a decrease in blood flow velocity although the volumetric flow may remain unchanged. The changes we have observed in the CRA may be secondary to dilation of the retinal vasculature. At present there is conflicting evidence as to whether blood flow velocity is increased or decreased in diabetic retinopathy. Different techniques and sites of measurement have contributed to the varying results. Using laser Doppler velocimetry Patel et al. (1992) showed an increase in blood flow velocity in temporal retinal veins in type 1 and 2 diabetics with untreated retinopathy compared to nondiabetic controls and diabetics with no retinopathy. In contrast, Feke et al. (1994) found that blood speeds were on average 33% lower using laser Doppler measurements from temporal retinal arteries in type 1 diabetics with NDR or BDR. Con-

A O S 2000 flicting data has also been published for macular blood flow. In the macular region the scanning laser technique has shown significantly reduced perifoveal capillary blood cell velocities with increased intercapillary areas (Arend et al. 1991) in type 1 and 2 diabetics with or without retinopathy. However, a study using blue field entopic stimulation found a 25% increase in macular capillary flow velocity in type 1 diabetics compared to controls (Sinclair 1991). The finding of slower velocites in the central retinal artery of diabetics, which become progressively slower with the severity of the retinopathy, is likely to have a significant effect on retinal perfusion and function. Harris et al. (1996) were able to show that hyperoxia improves contrast sensitivity in early diabetic retinopathy, therefore indicating that there is a link between tissue hypoxia and visual function in diabetics. Colour Doppler imaging gives a measure of velocity in blood vessels. As it is not possible to accurately measure the diameter of the vessels being studied with this technique, blood flow cannot be quantitatively calculated. How close the relationship between blood velocity and volume flow in the vessels studied is unknown. Our finding of reduced velocities in the CRA of diabetics should be interpreted with caution. As with many human vascular studies involving diabetics, vasoactive medication may become a confounding variable. Until it is possible to quantitatively measure blood flow in the ocular vasculature accurately, especially retinal tissue perfusion, the debate surrounding the pathogenesis of diabetic retinopathy will continue.

References Aburn NS & Sergott RC (1993): Orbital colour Doppler imaging. Eye 7: 639–647. Arend O, Wolf S, Jung F, Bertram B, Postgens H, Toonen H & Reim M (1991): Retinal microcirculation in patients with diabetes mellitus: dynamic and morphological analysis of perifoveal capillary network. Br J Ophthalmol 75: 514–518. Armitage P & Rose GA (1966): The variability of measurements of casual blood pressure. I. A laboratory study. Clin Sci 30: 325–335. Baxter GM & Williamson TH (1995): Colour Doppler imaging of the eye: normal ranges,

reproducibility and observer variation. J Ultrasound Med 14: 91–96. Butt Z, McKillop G, O’Brien C, Allan P & Aspinall P (1995): Measurement of ocular blood flow velocity using colour Doppler imaging in low tension glaucoma. Eye 9: 29–33. Diabetic retinopathy study research group (1981): A modification of the Airlie House classification of diabetic retinopathy: DRS report .7. Invest Ophthalmol Vis Sci 21: 210–226. Early treatment diabetic retinopathy study research group (1991): Grading diabetic retinopathy from stereoscopic color fundus photographs – an extension of the modified Airlie House classification: ETDRS Report .10 Ophthalmology 98: 786–806. Erickson SJ, Hendrix LE, Massaro BM, Harris GJ, Lewandowski MF, Foley WD & Lawson TL (1989): Colour Doppler flow imaging of the normal and abnormal orbit. Radiology 173: 511–516. Evans DW, Harris A, RP Danis, Arend O & Martin BJ (1997): Altered retrobulbar vascular reactivity in early diabetic retinopathy. Br J Ophthalmol 81: 279–282. Feke GT, Buzney SM, Ogasawara H, Fujio N, Goger DG, Spack NP & Gabbay KH (1994): Retinal circulatory abnormalities in type 1 diabetes. Invest Ophthalmol Vis Sci 35: 2968–2975. Forrester JV, Knott RM & McIntosh LC (1993): Pathogenesis of proliferative diabetic retinopathy and maculopathy. The Diabetes Annual 7: 178–191. Goebel W, Lieb WE, Ho A, Sergott RC, Farhoumand R & Grehn F (1995): Color Doppler imaging: a new technique to assess orbital blood flow in patients with diabetic retinopathy. Invest Ophthalmol Vis Sci 36: 864–870. Grunwald JE, DuPont J & Riva CE (1996): Retinal haemodynamics in patients with early diabetes mellitus. Br J Ophthalmol 80: 327–331. Harris A, Arend O, Danis R, Evans D, Wolf S & Martin B (1996): Hyperoxia improves contrast sensitivity in early diabetic retinopathy. Br J Ophthalmol 80: 209–213. Hart WM (Ed) (1992): Adler’s physiology of the eye, 9th edition. St Louis: Mosby Year Book. Kohner EM (1993): Diabetic retinopathy. Br Med J 307: 1195–1199. Kumar P & Clark M (Eds) (1998): Clinical Medicine, 4th edition. Edinburgh: WB Saunders. Lieb WE, Cohen SM, Merton DA, Shields JA, Mitchell DG & Goldberg BB (1991): Colour Doppler imaging of the eye and orbit. Technique and normal vascular anatomy. Arch Ophthalmol 109: 527–532. MacKinnon JR, O’Brien C, Swa K, Aspinall P, Butt Z & Cameron D (1997): Pulsatile

ocular blood flow in untreated diabetic retinopathy. Acta Ophthalmol Scand 75: 661– 664. McLeod DS & Lutty GA (1994): High resolution histologic analysis of the human choroidal vasculature. Invest Ophthalmol Vis Sci 35: 3799–3811. Mendivil A & Cuartero V (1996): Ocular blood flow velocities in patients with proliferative diabetic retinopathy after scatter photocoagulation. Retina 16: 222–227. Mendivil A, Cuartero V & Mendivil MP (1995): Ocular blood flow velocities in patients with proliferative diabetic retinopathy and healthy volunteers: a prospective study. Br J Ophthalmol 79: 413–416. Mulhern M, Hulsmans F, Geussen HC & Greve EL (1996): Resistivity indices in CDI of ocular vessels. Ophthalmology 103: 1333– 1334. Norusis MJ (1993): SPSS for windows. Base system user’s guide, Release 6.0. Chicago: SPSS Inc. Patel V, Rassam S, Newsom R, Wiek J & Kohner E (1992): Retinal blood flow in diabetic retinopathy. Br Med J 305: 678–683. Planiol T, Pourcelot L, Pottier JM & Degiovanni E (1972): Etude de la circulation carotidienne par les methodes ultrasoniques et la thermographie. Rev Neurol 126: 127–141. Rankin SJA, Walman BE, Buckley AR & Drance SM (1995): Colour Doppler imaging and spectral analysis of the optic nerve vasculature in glaucoma. Am J Ophthalmol 119: 685–693. Sinclair SH (1991): Macular retinal capillary hemodynamics in diabetic patients. Ophthalmology 98: 1580–1586. Ulbig MRW & Hamilton AMP (1993): Factors influencing the natural history of diabetic retinopathy. Eye 7: 242–249. Williamson TH, Lowe GD & Baxter GM (1995): Influence of age, systemic blood pressure, smoking and blood viscosity on orbital blood velocities. Br J Ophthalmol 79: 17–22. Winer BJ, Brown DR & Michels KM (1991): Statistical principles in experimental design, 3rd ed. New York: McGraw-Hill. Received on July 7th, 1998. Accepted on December 1st, 1999.

Corresponding author: Dr Jane R. MacKinnon Department of Ophthalmology Medical School Buildings Foresterhill Aberdeen AB9 2ZD. Scotland Tel: 01224 681 818, ext. 53215. Fax: 01224 840 746.

389