RETINAL VESSEL CALIBER CHANGES IN VASCULITIS GERALD LIEW, FRANZCO,*† ADNAN TUFAIL, FRCO,*‡ VICTORIA F. COSATTO, BMEDSCI,† AVA G. TAN, MPH,† JAVIER ZARRANZ-VENTURA, PHD,§ DAWN A. SIM, MBBS,*‡ PEARSE A. KEAN, PHD,*‡ CATHERINE A. EGAN, FRCOPHTH,*‡ PAUL MITCHELL, FRANZCO,† MARK C. WESTCOTT, FRCOPHTH,* RICHARD W. J. LEE, PHD, FRCOPHTH,*‡§ CARLOS E. PAVESIO, FRCOPHTH*‡ Background: Retinal vasculitis is a potentially sight-threatening inflammation of the retinal vessels, but little is known about the in vivo vascular changes, which occur in affected eyes. The authors therefore sought to measure vessel caliber in eyes with vasculitis. Methods: Retrospective case–control study. Vasculitis was confirmed using fluorescein angiography. Vessel calibers were measured using validated semiautomated software. Results: There were 21 eyes from 15 patients with vasculitis and 33 control eyes from 21 control subjects. Most cases were diagnosed with idiopathic vasculitis. All had periphlebitis, and one eye also had arteritis. After adjustment for age and gender, mean arteriolar caliber was 143 mm (95% confidence interval [CI], 134–152) in cases and 158 mm (95% CI, 151–165) in controls (P = 0.01). Venular caliber was similar in cases (229 mm; 95% CI, 215– 243) and controls (228 mm; 95% CI, 217–234; P = 0.91), whereas arteriole-to-venule ratio was smaller in cases (0.63; 95% CI, 0.60–0.66) compared with controls (0.70; 95% CI, 0.02–0.11; P = 0.004). Conclusion: Retinal vasculitis was associated with narrower arteriolar caliber, whereas venular caliber was similar to controls. This resulted in a smaller arteriole-to-venule ratio in eyes with vasculitis. RETINA 35:803–808, 2015
R
losis),1–5 whereas systemic inflammatory conditions include Behcet, sarcoidosis, and systemic lupus erythematosis.1–3,6 Inflammation involving retinal venules (periphlebitis) is more commonly observed than inflammation of the arterioles (arteritis). Histopathologic specimens show in general an intramural lymphocytic infiltrate of the vessel walls, and in severe cases, retinal exudates are often present.7–9 The presence of retinal vasculitis is usually detected clinically where it presents as confluent or segmental (skip lesions) sheathing or cuffing of the retinal vessels. Complications of retinal vasculitis include occlusion of the vessels with associated retinal hemorrhages and ischemic cotton-wool spots, retinal hard exudates, and cystoid macular edema. Fluorescein angiography is often used to confirm the diagnosis and can detect areas of involvement that are not evident clinically. Characteristic features seen with fluorescein angiography in vasculitis include staining of the vessel wall with fluorescein due to breakdown of the inner blood retinal barrier and leakage into surrounding retinal tissue. Nonetheless, little is known
etinal vasculitis refers to inflammation of the retinal vessel walls and can occur either in isolation or in association with ocular and systemic diseases. Ocular diseases associated with retinal vasculitis include infections (e.g., acute retinal necrosis, tubercu-
From the *Moorfields Eye Hospital NHS Foundation Trust, London, United Kingdom; †Center for Vision Research, Westmead Millennium Institute, University of Sydney, Sydney, Australia; ‡Institute of Ophthalmology, University College London, London, United Kingdom; and §Bristol Eye Hospital, University Hospitals Bristol NHS Foundation Trust, Bristol, United Kingdom. Supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology. G. Liew received funding from the RANZCO Eye Foundation. D. A. Sim received funding from Fight For Sight UK, Grant 1987. The remaining authors have no conflicting interests to disclose. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.retinajournal.com). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. Reprint requests: Gerald Liew, FRANZCO, Center for Vision Research, Westmead Millennium Institute, Hawkesbury Road, NSW 2145, Sydney, Australia; e-mail: [email protected]
803
804
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 4
about other aspects of vasculitis, such as whether the inflamed vessels dilate or constrict, which would have implications for perfusion and blood flow. We therefore conducted a case–control study to determine invivo retinal vessel caliber in eyes with vasculitis. Methods This study is a retrospective hospital-based case–control study of patients presenting to retinal clinics at Moorfields Eye Hospital National Health Service Foundation Trust, London, United Kingdom over a 6-month period from January 1, 2012, to June 30, 2012. All cases with vasculitis were confirmed using fundus fluorescein angiography. Patients were started on immunosuppressive treatment after baseline fundus photography and investigations—hence, no patients were on immunosuppressive therapy at the time of vessel caliber measurement. Controls were identified from attendees to a general primary care outpatient eye clinic at Moorfields Eye Hospital over the same period, most of whom had incidental findings of choroidal nevi. Approval for data collection and analysis was obtained from the local ethics committee and adhered to the tenets set forth in the Declaration of Helsinki. Patient demographic data, consultant confirmed diagnosis, and treatments delivered were obtained from electronic patient records and patient chart review. Fundus photographs and fluorescein angiograms were obtained and manually reviewed to confirm the diagnosis, and the best quality color fundus images were anonymized and electronically transferred to the Center for Vision Research, Sydney, Australia, for retinal vessel caliber quantification by a single trained grader masked to patient diagnosis and other patient data. Quantification of Retinal Vessel Calibers All photographic images were acquired with a digital retinal camera system (Topcon TRC 50IX; Topcon Medical Systems Inc, Paramus, NJ). A computerassisted grading method with high reproducibility was used to measure retinal vessel caliber.10 This method to quantitatively and reproducibly measure arteriolar and venular caliber from digitized retinal photographs has been validated and applied in numerous studies.11–14 These studies showed that small changes in retinal arteriolar and venular caliber can be reliably measured and carry prognostic information that predicts the risk of cardiovascular disease.11–14 Further details are published elsewhere.10,15 In brief, digitized fundus images were displayed, and all vessels greater than 25 mm in diameter and completely passing through
the region between 1/2 to 1 disk diameter from the optic disk margin were measured (Figure 1). Average arteriolar and venular calibers were then calculated,10 and the arteriole-to-venule ratio (AVR) was defined as the ratio of mean arteriolar to venular calibers. Reproducibility is high, with intragrader and intergrader kappas of 0.88 to 0.93 and 0.82 to 0.90, respectively.10 For inclusion in the study, all image sets had to be of sufficient quality to allow visualization of arterioles and venules exiting from the disk to allow measurement of arteriolar and venular caliber. Statistical Analysis Statistical analyses were performed using SAS version 9 (SAS Institute Inc, Cary, NC). Chi-square tests were performed to examine differences between categorical variables for cases and controls, and mixed models were constructed to examine the association of vessel caliber with the presence of vasculitis compared with controls. Multivariable models were adjusted for age (continuous), gender (male, female), and camera magnification (Camera 1, Camera 2) as terms in the model. We report unadjusted vessel calibers in cases and controls, adjusted calibers, P values, and 95% confidence intervals (CIs). Results We analyzed 21 eyes from 15 patients with vasculitis and 33 control eyes from 21 control subjects that met our image quality inclusion criteria (Table 1). The mean age of cases was 46.8 ± 18.7 (SD) years, 52% were female, and 48% of eyes included were right eyes. Control eyes were similar with regard to these characteristics. Mean arteriolar caliber was narrower in cases than in controls (143 mm; 95% CI, 134– 152 and 158 mm; 95% CI, 152–165, respectively; P , 0.01), whereas venules were of similar caliber (229 mm; 95% CI, 212–246 and 228 mm; 95% CI, 219–237, respectively; P = 0.90). As a result of the arteriolar narrowing, AVR in cases was smaller at 0.63 (95% CI, 0.59–0.67) compared with 0.70 (95% CI, 0.68–0.72) in controls (P , 0.01), a 10% reduction (Figure 2). The arteriolar narrowing can be appreciated in Figure 1, A–C, compared with a normal control (Figure 1D). Adjustment for age and gender did not change these findings appreciably, with a mean adjusted arteriolar caliber of 143 mm (95% CI, 134– 152) in cases and 158 mm (95% CI, 151–165) in controls (P = 0.01) (Table 2). Venular caliber remained similar in cases and controls, although AVR was smaller in cases (0.63; 95% CI, 0.60–0.66) compared with controls (0.70; 95% CI, 0.02–0.11; P = 0.004).
VESSEL CALIBER AND VASCULITIS LIEW ET AL
805
Fig. 1. Color photographs, fundus fluorescein angiograms, and retinal vessel caliber measurements. A–C. Patients with idiopathic vasculitis, vasculitis associated with tuberculosis, and vasculitis associated with birdshot choroidopathy. The arrows show venules, whereas the arrowheads show arterioles, which are narrowed compared with the venules. D. Patient with a nevus (normal control) showing the normal relationship of arteriolar and venular calibers.
We performed further supplementary analyses where these were restricted by eye and obtained similar results. For example, when analyses were performed in only right eyes (10 cases and 17 controls), mean arteriolar caliber in cases was 142 mm (95% CI, 126–158) compared with 161 mm (95% CI, 150–172;
P = 0.04) in controls, and venular caliber remained similar in cases and controls (228 mm; 95% CI, 196–259 and 226 mm; 95% CI, 214–238, respectively; P = 0.93). Excluding the two eyes without clinically active vasculitis also did not change the results. Three patients (five eyes: four with idiopathic vasculitis
806
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 4
Table 1. Characteristics and Diagnoses of the Patients and Eyes Included in the Study Cases Controls (n, Eyes = 21) (n, Eyes = 33) Patients, n Mean age (SD), years Female, n (%) Right eyes, n (%) Diagnoses Idiopathic periphlebitis Tuberculosis-related periphlebitis Birdshot periphlebitis Behcet periphlebitis Systemic lupus erythematosis-related periphlebitis Idiopathic periphlebitis and arteritis Anterior uveitis Nevi Central serous chorioretinopathy Cataract Achromatopsia Retinal hole (round) Normal
15 46.8 (18.7) 11 (52) 10 (48)
21 43.2 (19.0) 22 (67) 17 (52)
9 5
— —
2 2 2
— — —
1
—
— — —
3 5 2
— — — —
1 1 1 20
and one with tuberculous vasculitis) had evidence of retinal ischemia on their fluorescein angiograms. Excluding these patients from analysis did not alter the findings.
Fig. 2. Arteriole-to-venule ratio in cases (eyes) with vasculitis and controls. Error bars are SDs.
We also performed supplementary analyses by camera (Supplemental Digital Content 1, http:// links.lww.com/IAE/A317). When analyses were restricted to images taken using Camera 1 (n = 30), arteriolar caliber remained smaller in cases than controls (145 mm; 95% CI, 136–155 and 156 mm; 95% CI, 146–166), but this difference was not significant (P = 0.12). Venular caliber remained similar in both cases and controls, and AVR was significantly smaller in cases than controls (0.63; 95%, 95% CI, 0.59–0.67 and 0.69; 95% CI, 0.65–0.74, respectively; P = 0.03). Analyses restricted to images taken from Camera 2 (n = 24) produced similar results, with only AVR significantly smaller in cases compared with controls (0.63; 95% CI, 0.57–0.69 and 0.70; 95% CI, 0.67–0.73, respectively; P = 0.04). Discussion Retinal vasculitis is an inflammatory condition affecting the retinal vessels that can be sight threatening in severe cases. We report preliminary findings from a case–control study that arterioles in eyes with vasculitis are on average 15 mm narrower than in eyes without vasculitis. Venules in these eyes were of similar caliber, and AVR was reduced to 0.63 in cases as compared with a more normal 0.70 in controls. There are few studies with which to compare our results, as to our knowledge, in vivo measurement of retinal vessel calibers in patients with vasculitis has not been previously performed. The limited number of histopathologic studies of retinal vasculitis that do exist have not reported retinal vessel calibers, which are likely in any case to be inaccurate because of loss of hydrostatic pressure and tissue processing. Perivascular and intramural lymphocytic infiltration of the retinal vessels has been reported in patients with both Birdshot choriodopathy9 and Behcet disease,16 whereas nodular proliferation of epithelioid cells has been reported in a patient with sarcoidosis.17 Takayasu disease, a vasculitis that affects large arteries, results in transmural inflammation and subsequent arterial stenosis in some segments and dilation in others.18 Our finding of arteriolar narrowing, without venular caliber changes, is of interest because in all but one patient, only the venules were inflamed clinically and angiographically whereas the arterioles seemed unaffected. Furthermore, our measurements of vessel caliber were taken at a peripapillary distance of one half to one disk diameter away from the optic disk, a location considerably remote from most areas of involved vasculitis on clinical examination and on angiography, yet we were able to detect a generalized reduction in arteriolar
807
VESSEL CALIBER AND VASCULITIS LIEW ET AL Table 2. Age- and Sex-Adjusted Retinal Vessel Calibers in Cases With Vasculitis Compared With Control Eyes Age- and Sex-Adjusted Caliber (95% CI), mm Vessel Measurement Arteriolar caliber Venular caliber AVR
Cases With Vasculitis (n = 21)
Controls Without Vasculitis (n = 33)
Difference
P
143 (134–152) 229 (215–243) 0.63 (0.60–0.66)
158 (151–165) 228 (217–234) 0.70 (0.67–0.72)
15 (4–27) −1 (−19–17) 0.07 (0.02–0.11)
0.01 0.91 0.004
caliber. A possible explanation for the findings may be that some of the inflammatory cytokines associated with vasculitis also trigger generalized arteriolar vasoconstriction. For example, tumor necrosis factor-a is an important proinflammatory cytokine that plays a major role in Behcet disease19 and other vasculitides20 and been shown to have vasoconstrictive effects on arterioles.21,22 Nonetheless, most of our patients did not have Behcet disease, and it is unclear if this cytokine plays a role in this study. Alternatively, retinal ischemia from vascular occlusion could potentially account for some of the findings, although excluding the five eyes with angiographic evidence of retinal ischemia did not change the results. One limitation that should be borne in mind when interpreting these results is that we did not perform ultra-widefield imaging, which may have detected areas of peripheral ischemia not visible on standard 30° imaging. The presence of large regions of peripheral ischemia could potentially explain the arteriolar narrowing we observed. Future studies should examine ultra-widefield changes in vasculitis, and if peripheral ischemia is detected, if this is related to arteriolar caliber changes. All except two patients in this study had bilateral vasculitis; the two patients with unilateral vasculitis had idiopathic vasculitis. We could not compare involved and uninvolved eyes, as the photographs of the two uninvolved eyes were of insufficient quality to permit accurate grading. This study, although small, provides proof-of-concept that eyes with vasculitis have narrower arteriolar caliber, which may be detected clinically as a small reduction in the AVR (Figure 2). It should be borne in mind that other conditions, namely elevated blood pressure and aging, are also associated with narrowed arterioles. Of more interest is the possibility that vessel caliber changes may be a marker of vasculitic activity or response to therapy. Longitudinal follow-up of this and other cohorts is warranted to determine the prognostic significance of these arteriolar caliber changes and if they reverse with control of inflammation. Strengths of this study include careful selection of patients with confirmed vasculitis clinically and angiographically, their assessment before initiation of therapy, assessment of vascular calibers with validated semiautomated software, and adjustment for confound-
ers such as age. Limitations include the lack of information on confounders such as the duration of elevated blood pressure and medications and the restricted field of analysis around the disk, which may not adequately reflect peripheral vessel changes. Another limitation is the small sample size. A potential limitation is that retinal photography was not gated to the cardiac cycle. Individual retinal vessel calibers have been found to vary with cardiac cycle; however, the summary measure of mean vessel caliber that our program measures does not vary significantly with cardiac cycle23 and has been found to be stable to shortterm influences such as week-to-week variation in blood pressure.24 Finally, a moderate number of images were of inadequate quality for grading (9 eyes of 30 in cases and 9 of 42 in controls). Although we think that this is purely related to technical factors in the control group, this may nonetheless have introduced a selection bias, which influenced our results. As some of the images were ungradable because of vitritis in patients with vasculitis, which may indicate more severe inflammation and more arteriolar narrowing, it is also possible that our results may underestimate the degree of narrowing in vasculitis. This is a weakness that limits the applicability of retinal vessel measurement as a marker of vasculitis. For this study, we excluded images with an image quality that was worse than a set standard, which excluded cases with more severe vitritis. Poor-quality images may lead to overestimates of arteriolar caliber and may have occurred in this study.25 This would not alter our findings, as we report that narrower arterioles are associated with vasculitis and may rather have led to an underestimate of the arteriolar narrowing. In summary, this retrospective case–control study found that systemic vasculitis was associated with narrower arteriolar caliber, while venular caliber remained unchanged. This resulted in a smaller AVR in eyes with vasculitis, which was detectable using standard digital retinal images. The results suggest that generalized retinal vascular changes may occur in affected patients and that a smaller AVR may be a useful clinical sign of vasculitis. Key words: retinal vessel, vasculitis, arteritis, periphlebitis, inflammation.
808
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 4
References 1. Abu El-Asrar AM, Herbort CP, Tabbara KF. Retinal vasculitis. Ocul Immunol Inflamm 2005;13:415–433. 2. Graham EM, Stanford MR, Sanders MD, et al. A point prevalence study of 150 patients with idiopathic retinal vasculitis: 1. Diagnostic value of ophthalmological features. Br J Ophthalmol 1989;73:714–721. 3. El Asrar AM, Herbort CP, Tabbara KF. A clinical approach to the diagnosis of retinal vasculitis. Int Ophthalmol 2010;30: 149–173. 4. Gupta A, Gupta V, Arora S, et al. PCR-positive tubercular retinal vasculitis: clinical characteristics and management. Retina 2001;21:435–444. 5. Fountain JA, Werner RB. Tuberculous retinal vasculitis. Retina 1984;4:48–50. 6. Neumann R, Foster CS. Corticosteroid-sparing strategies in the treatment of retinal vasculitis in systemic lupus erythematosus. Retina 1995;15:201–212. 7. Gass JD, Olson CL. Sarcoidosis with optic nerve and retinal involvement. Arch Ophthalmol 1976;94:945–950. 8. Pederson JE, Kenyon KR, Green WR, Maumenee AE. Pathology of pars planitis. Am J Ophthalmol 1978;86:762–774. 9. Gaudio PA, Kaye DB, Crawford JB. Histopathology of birdshot retinochoroidopathy. Br J Ophthalmol 2002;86:1439–1441. 10. Sherry LM, Wang JJ, Rochtchina E, et al. Reliability of computer-assisted retinal vessel measurementin a population. Clin Exp Ophthalmol 2002;30:179–182. 11. Liew G, Wang JJ, Mitchell P, Wong TY. Retinal vascular imaging: a new tool in microvascular disease research. Circ Cardiovasc Imaging 2008;1:156–161. 12. Retinal pigment epithelial detachments in the elderly: a controlled trial of argon laser photocoagulation. Br J Ophthalmol 1982;66:1–16. 13. Wong TY, Mitchell P. Hypertensive retinopathy. N Engl J Med 2004;351:2310–2317.
14. McGeechan K, Liew G, Macaskill P, et al. Meta-analysis: retinal vessel caliber and risk for coronary heart disease. Ann Intern Med 2009;151:404–413. 15. Gopinath B, Wang JJ, Kifley A, et al. The association between ocular biometry and retinal vascular caliber is comparable from early childhood to adolescence. Invest Ophthalmol Vis Sci 2013;54:1501–1508. 16. Charteris DG, Champ C, Rosenthal AR, Lightman SL. Behcet’s disease: activated T lymphocytes in retinal perivasculitis. Br J Ophthalmol 1992;76:499–501. 17. Klein BE, Klein R, Jensen SC, Ritter LL. Are sex hormones associated with age-related maculopathy in women? The Beaver Dam Eye Study. Trans Am Ophthalmol Soc 1994;92: 289–295. 18. Vaideeswar P, Deshpande JR. Pathology of Takayasu arteritis: a brief review. Ann Pediatr Cardiol 2013;6:52–58. 19. Sugita S, Kawazoe Y, Imai A, et al. Role of IL-22- and TNFalpha-producing Th22 cells in uveitis patients with Behcet’s disease. J Immunol 2013;190:5799–5808. 20. Cruz BA, Reis DD, Araujo CA. Refractory retinal vasculitis due to sarcoidosis successfully treated with infliximab. Rheumatol Int 2007;27:1181–1183. 21. Vila E, Salaices M. Cytokines and vascular reactivity in resistance arteries. Am J Physiol Heart Circ Physiol 2005;288: H1016–H1021. 22. Wagner EM. TNF-alpha induced bronchial vasoconstriction. Am J Physiol Heart Circ Physiol 2000;279:H946–H951. 23. Hao H, Sasongko MB, Wong TY, et al. Does retinal vascular geometry vary with cardiac cycle? Invest Ophthalmol Vis Sci 2012;53:5799–5805. 24. McCanna CD, Myers CE, Lee M, et al. Variability of measurement of retinal vessel diameters. Ophthalmic Epidemiol 2013; 20:392–401. 25. Chandler CS, Gangaputra S, Hubbard LD, et al. Suboptimal image focus broadens retinal vessel caliber measurement. Invest Ophthalmol Vis Sci 2011;52:8558–8561.
R
losis),1–5 whereas systemic inflammatory conditions include Behcet, sarcoidosis, and systemic lupus erythematosis.1–3,6 Inflammation involving retinal venules (periphlebitis) is more commonly observed than inflammation of the arterioles (arteritis). Histopathologic specimens show in general an intramural lymphocytic infiltrate of the vessel walls, and in severe cases, retinal exudates are often present.7–9 The presence of retinal vasculitis is usually detected clinically where it presents as confluent or segmental (skip lesions) sheathing or cuffing of the retinal vessels. Complications of retinal vasculitis include occlusion of the vessels with associated retinal hemorrhages and ischemic cotton-wool spots, retinal hard exudates, and cystoid macular edema. Fluorescein angiography is often used to confirm the diagnosis and can detect areas of involvement that are not evident clinically. Characteristic features seen with fluorescein angiography in vasculitis include staining of the vessel wall with fluorescein due to breakdown of the inner blood retinal barrier and leakage into surrounding retinal tissue. Nonetheless, little is known
etinal vasculitis refers to inflammation of the retinal vessel walls and can occur either in isolation or in association with ocular and systemic diseases. Ocular diseases associated with retinal vasculitis include infections (e.g., acute retinal necrosis, tubercu-
From the *Moorfields Eye Hospital NHS Foundation Trust, London, United Kingdom; †Center for Vision Research, Westmead Millennium Institute, University of Sydney, Sydney, Australia; ‡Institute of Ophthalmology, University College London, London, United Kingdom; and §Bristol Eye Hospital, University Hospitals Bristol NHS Foundation Trust, Bristol, United Kingdom. Supported by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology. G. Liew received funding from the RANZCO Eye Foundation. D. A. Sim received funding from Fight For Sight UK, Grant 1987. The remaining authors have no conflicting interests to disclose. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.retinajournal.com). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. Reprint requests: Gerald Liew, FRANZCO, Center for Vision Research, Westmead Millennium Institute, Hawkesbury Road, NSW 2145, Sydney, Australia; e-mail: [email protected]
803
804
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 4
about other aspects of vasculitis, such as whether the inflamed vessels dilate or constrict, which would have implications for perfusion and blood flow. We therefore conducted a case–control study to determine invivo retinal vessel caliber in eyes with vasculitis. Methods This study is a retrospective hospital-based case–control study of patients presenting to retinal clinics at Moorfields Eye Hospital National Health Service Foundation Trust, London, United Kingdom over a 6-month period from January 1, 2012, to June 30, 2012. All cases with vasculitis were confirmed using fundus fluorescein angiography. Patients were started on immunosuppressive treatment after baseline fundus photography and investigations—hence, no patients were on immunosuppressive therapy at the time of vessel caliber measurement. Controls were identified from attendees to a general primary care outpatient eye clinic at Moorfields Eye Hospital over the same period, most of whom had incidental findings of choroidal nevi. Approval for data collection and analysis was obtained from the local ethics committee and adhered to the tenets set forth in the Declaration of Helsinki. Patient demographic data, consultant confirmed diagnosis, and treatments delivered were obtained from electronic patient records and patient chart review. Fundus photographs and fluorescein angiograms were obtained and manually reviewed to confirm the diagnosis, and the best quality color fundus images were anonymized and electronically transferred to the Center for Vision Research, Sydney, Australia, for retinal vessel caliber quantification by a single trained grader masked to patient diagnosis and other patient data. Quantification of Retinal Vessel Calibers All photographic images were acquired with a digital retinal camera system (Topcon TRC 50IX; Topcon Medical Systems Inc, Paramus, NJ). A computerassisted grading method with high reproducibility was used to measure retinal vessel caliber.10 This method to quantitatively and reproducibly measure arteriolar and venular caliber from digitized retinal photographs has been validated and applied in numerous studies.11–14 These studies showed that small changes in retinal arteriolar and venular caliber can be reliably measured and carry prognostic information that predicts the risk of cardiovascular disease.11–14 Further details are published elsewhere.10,15 In brief, digitized fundus images were displayed, and all vessels greater than 25 mm in diameter and completely passing through
the region between 1/2 to 1 disk diameter from the optic disk margin were measured (Figure 1). Average arteriolar and venular calibers were then calculated,10 and the arteriole-to-venule ratio (AVR) was defined as the ratio of mean arteriolar to venular calibers. Reproducibility is high, with intragrader and intergrader kappas of 0.88 to 0.93 and 0.82 to 0.90, respectively.10 For inclusion in the study, all image sets had to be of sufficient quality to allow visualization of arterioles and venules exiting from the disk to allow measurement of arteriolar and venular caliber. Statistical Analysis Statistical analyses were performed using SAS version 9 (SAS Institute Inc, Cary, NC). Chi-square tests were performed to examine differences between categorical variables for cases and controls, and mixed models were constructed to examine the association of vessel caliber with the presence of vasculitis compared with controls. Multivariable models were adjusted for age (continuous), gender (male, female), and camera magnification (Camera 1, Camera 2) as terms in the model. We report unadjusted vessel calibers in cases and controls, adjusted calibers, P values, and 95% confidence intervals (CIs). Results We analyzed 21 eyes from 15 patients with vasculitis and 33 control eyes from 21 control subjects that met our image quality inclusion criteria (Table 1). The mean age of cases was 46.8 ± 18.7 (SD) years, 52% were female, and 48% of eyes included were right eyes. Control eyes were similar with regard to these characteristics. Mean arteriolar caliber was narrower in cases than in controls (143 mm; 95% CI, 134– 152 and 158 mm; 95% CI, 152–165, respectively; P , 0.01), whereas venules were of similar caliber (229 mm; 95% CI, 212–246 and 228 mm; 95% CI, 219–237, respectively; P = 0.90). As a result of the arteriolar narrowing, AVR in cases was smaller at 0.63 (95% CI, 0.59–0.67) compared with 0.70 (95% CI, 0.68–0.72) in controls (P , 0.01), a 10% reduction (Figure 2). The arteriolar narrowing can be appreciated in Figure 1, A–C, compared with a normal control (Figure 1D). Adjustment for age and gender did not change these findings appreciably, with a mean adjusted arteriolar caliber of 143 mm (95% CI, 134– 152) in cases and 158 mm (95% CI, 151–165) in controls (P = 0.01) (Table 2). Venular caliber remained similar in cases and controls, although AVR was smaller in cases (0.63; 95% CI, 0.60–0.66) compared with controls (0.70; 95% CI, 0.02–0.11; P = 0.004).
VESSEL CALIBER AND VASCULITIS LIEW ET AL
805
Fig. 1. Color photographs, fundus fluorescein angiograms, and retinal vessel caliber measurements. A–C. Patients with idiopathic vasculitis, vasculitis associated with tuberculosis, and vasculitis associated with birdshot choroidopathy. The arrows show venules, whereas the arrowheads show arterioles, which are narrowed compared with the venules. D. Patient with a nevus (normal control) showing the normal relationship of arteriolar and venular calibers.
We performed further supplementary analyses where these were restricted by eye and obtained similar results. For example, when analyses were performed in only right eyes (10 cases and 17 controls), mean arteriolar caliber in cases was 142 mm (95% CI, 126–158) compared with 161 mm (95% CI, 150–172;
P = 0.04) in controls, and venular caliber remained similar in cases and controls (228 mm; 95% CI, 196–259 and 226 mm; 95% CI, 214–238, respectively; P = 0.93). Excluding the two eyes without clinically active vasculitis also did not change the results. Three patients (five eyes: four with idiopathic vasculitis
806
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 4
Table 1. Characteristics and Diagnoses of the Patients and Eyes Included in the Study Cases Controls (n, Eyes = 21) (n, Eyes = 33) Patients, n Mean age (SD), years Female, n (%) Right eyes, n (%) Diagnoses Idiopathic periphlebitis Tuberculosis-related periphlebitis Birdshot periphlebitis Behcet periphlebitis Systemic lupus erythematosis-related periphlebitis Idiopathic periphlebitis and arteritis Anterior uveitis Nevi Central serous chorioretinopathy Cataract Achromatopsia Retinal hole (round) Normal
15 46.8 (18.7) 11 (52) 10 (48)
21 43.2 (19.0) 22 (67) 17 (52)
9 5
— —
2 2 2
— — —
1
—
— — —
3 5 2
— — — —
1 1 1 20
and one with tuberculous vasculitis) had evidence of retinal ischemia on their fluorescein angiograms. Excluding these patients from analysis did not alter the findings.
Fig. 2. Arteriole-to-venule ratio in cases (eyes) with vasculitis and controls. Error bars are SDs.
We also performed supplementary analyses by camera (Supplemental Digital Content 1, http:// links.lww.com/IAE/A317). When analyses were restricted to images taken using Camera 1 (n = 30), arteriolar caliber remained smaller in cases than controls (145 mm; 95% CI, 136–155 and 156 mm; 95% CI, 146–166), but this difference was not significant (P = 0.12). Venular caliber remained similar in both cases and controls, and AVR was significantly smaller in cases than controls (0.63; 95%, 95% CI, 0.59–0.67 and 0.69; 95% CI, 0.65–0.74, respectively; P = 0.03). Analyses restricted to images taken from Camera 2 (n = 24) produced similar results, with only AVR significantly smaller in cases compared with controls (0.63; 95% CI, 0.57–0.69 and 0.70; 95% CI, 0.67–0.73, respectively; P = 0.04). Discussion Retinal vasculitis is an inflammatory condition affecting the retinal vessels that can be sight threatening in severe cases. We report preliminary findings from a case–control study that arterioles in eyes with vasculitis are on average 15 mm narrower than in eyes without vasculitis. Venules in these eyes were of similar caliber, and AVR was reduced to 0.63 in cases as compared with a more normal 0.70 in controls. There are few studies with which to compare our results, as to our knowledge, in vivo measurement of retinal vessel calibers in patients with vasculitis has not been previously performed. The limited number of histopathologic studies of retinal vasculitis that do exist have not reported retinal vessel calibers, which are likely in any case to be inaccurate because of loss of hydrostatic pressure and tissue processing. Perivascular and intramural lymphocytic infiltration of the retinal vessels has been reported in patients with both Birdshot choriodopathy9 and Behcet disease,16 whereas nodular proliferation of epithelioid cells has been reported in a patient with sarcoidosis.17 Takayasu disease, a vasculitis that affects large arteries, results in transmural inflammation and subsequent arterial stenosis in some segments and dilation in others.18 Our finding of arteriolar narrowing, without venular caliber changes, is of interest because in all but one patient, only the venules were inflamed clinically and angiographically whereas the arterioles seemed unaffected. Furthermore, our measurements of vessel caliber were taken at a peripapillary distance of one half to one disk diameter away from the optic disk, a location considerably remote from most areas of involved vasculitis on clinical examination and on angiography, yet we were able to detect a generalized reduction in arteriolar
807
VESSEL CALIBER AND VASCULITIS LIEW ET AL Table 2. Age- and Sex-Adjusted Retinal Vessel Calibers in Cases With Vasculitis Compared With Control Eyes Age- and Sex-Adjusted Caliber (95% CI), mm Vessel Measurement Arteriolar caliber Venular caliber AVR
Cases With Vasculitis (n = 21)
Controls Without Vasculitis (n = 33)
Difference
P
143 (134–152) 229 (215–243) 0.63 (0.60–0.66)
158 (151–165) 228 (217–234) 0.70 (0.67–0.72)
15 (4–27) −1 (−19–17) 0.07 (0.02–0.11)
0.01 0.91 0.004
caliber. A possible explanation for the findings may be that some of the inflammatory cytokines associated with vasculitis also trigger generalized arteriolar vasoconstriction. For example, tumor necrosis factor-a is an important proinflammatory cytokine that plays a major role in Behcet disease19 and other vasculitides20 and been shown to have vasoconstrictive effects on arterioles.21,22 Nonetheless, most of our patients did not have Behcet disease, and it is unclear if this cytokine plays a role in this study. Alternatively, retinal ischemia from vascular occlusion could potentially account for some of the findings, although excluding the five eyes with angiographic evidence of retinal ischemia did not change the results. One limitation that should be borne in mind when interpreting these results is that we did not perform ultra-widefield imaging, which may have detected areas of peripheral ischemia not visible on standard 30° imaging. The presence of large regions of peripheral ischemia could potentially explain the arteriolar narrowing we observed. Future studies should examine ultra-widefield changes in vasculitis, and if peripheral ischemia is detected, if this is related to arteriolar caliber changes. All except two patients in this study had bilateral vasculitis; the two patients with unilateral vasculitis had idiopathic vasculitis. We could not compare involved and uninvolved eyes, as the photographs of the two uninvolved eyes were of insufficient quality to permit accurate grading. This study, although small, provides proof-of-concept that eyes with vasculitis have narrower arteriolar caliber, which may be detected clinically as a small reduction in the AVR (Figure 2). It should be borne in mind that other conditions, namely elevated blood pressure and aging, are also associated with narrowed arterioles. Of more interest is the possibility that vessel caliber changes may be a marker of vasculitic activity or response to therapy. Longitudinal follow-up of this and other cohorts is warranted to determine the prognostic significance of these arteriolar caliber changes and if they reverse with control of inflammation. Strengths of this study include careful selection of patients with confirmed vasculitis clinically and angiographically, their assessment before initiation of therapy, assessment of vascular calibers with validated semiautomated software, and adjustment for confound-
ers such as age. Limitations include the lack of information on confounders such as the duration of elevated blood pressure and medications and the restricted field of analysis around the disk, which may not adequately reflect peripheral vessel changes. Another limitation is the small sample size. A potential limitation is that retinal photography was not gated to the cardiac cycle. Individual retinal vessel calibers have been found to vary with cardiac cycle; however, the summary measure of mean vessel caliber that our program measures does not vary significantly with cardiac cycle23 and has been found to be stable to shortterm influences such as week-to-week variation in blood pressure.24 Finally, a moderate number of images were of inadequate quality for grading (9 eyes of 30 in cases and 9 of 42 in controls). Although we think that this is purely related to technical factors in the control group, this may nonetheless have introduced a selection bias, which influenced our results. As some of the images were ungradable because of vitritis in patients with vasculitis, which may indicate more severe inflammation and more arteriolar narrowing, it is also possible that our results may underestimate the degree of narrowing in vasculitis. This is a weakness that limits the applicability of retinal vessel measurement as a marker of vasculitis. For this study, we excluded images with an image quality that was worse than a set standard, which excluded cases with more severe vitritis. Poor-quality images may lead to overestimates of arteriolar caliber and may have occurred in this study.25 This would not alter our findings, as we report that narrower arterioles are associated with vasculitis and may rather have led to an underestimate of the arteriolar narrowing. In summary, this retrospective case–control study found that systemic vasculitis was associated with narrower arteriolar caliber, while venular caliber remained unchanged. This resulted in a smaller AVR in eyes with vasculitis, which was detectable using standard digital retinal images. The results suggest that generalized retinal vascular changes may occur in affected patients and that a smaller AVR may be a useful clinical sign of vasculitis. Key words: retinal vessel, vasculitis, arteritis, periphlebitis, inflammation.
808
RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2015 VOLUME 35 NUMBER 4
References 1. Abu El-Asrar AM, Herbort CP, Tabbara KF. Retinal vasculitis. Ocul Immunol Inflamm 2005;13:415–433. 2. Graham EM, Stanford MR, Sanders MD, et al. A point prevalence study of 150 patients with idiopathic retinal vasculitis: 1. Diagnostic value of ophthalmological features. Br J Ophthalmol 1989;73:714–721. 3. El Asrar AM, Herbort CP, Tabbara KF. A clinical approach to the diagnosis of retinal vasculitis. Int Ophthalmol 2010;30: 149–173. 4. Gupta A, Gupta V, Arora S, et al. PCR-positive tubercular retinal vasculitis: clinical characteristics and management. Retina 2001;21:435–444. 5. Fountain JA, Werner RB. Tuberculous retinal vasculitis. Retina 1984;4:48–50. 6. Neumann R, Foster CS. Corticosteroid-sparing strategies in the treatment of retinal vasculitis in systemic lupus erythematosus. Retina 1995;15:201–212. 7. Gass JD, Olson CL. Sarcoidosis with optic nerve and retinal involvement. Arch Ophthalmol 1976;94:945–950. 8. Pederson JE, Kenyon KR, Green WR, Maumenee AE. Pathology of pars planitis. Am J Ophthalmol 1978;86:762–774. 9. Gaudio PA, Kaye DB, Crawford JB. Histopathology of birdshot retinochoroidopathy. Br J Ophthalmol 2002;86:1439–1441. 10. Sherry LM, Wang JJ, Rochtchina E, et al. Reliability of computer-assisted retinal vessel measurementin a population. Clin Exp Ophthalmol 2002;30:179–182. 11. Liew G, Wang JJ, Mitchell P, Wong TY. Retinal vascular imaging: a new tool in microvascular disease research. Circ Cardiovasc Imaging 2008;1:156–161. 12. Retinal pigment epithelial detachments in the elderly: a controlled trial of argon laser photocoagulation. Br J Ophthalmol 1982;66:1–16. 13. Wong TY, Mitchell P. Hypertensive retinopathy. N Engl J Med 2004;351:2310–2317.
14. McGeechan K, Liew G, Macaskill P, et al. Meta-analysis: retinal vessel caliber and risk for coronary heart disease. Ann Intern Med 2009;151:404–413. 15. Gopinath B, Wang JJ, Kifley A, et al. The association between ocular biometry and retinal vascular caliber is comparable from early childhood to adolescence. Invest Ophthalmol Vis Sci 2013;54:1501–1508. 16. Charteris DG, Champ C, Rosenthal AR, Lightman SL. Behcet’s disease: activated T lymphocytes in retinal perivasculitis. Br J Ophthalmol 1992;76:499–501. 17. Klein BE, Klein R, Jensen SC, Ritter LL. Are sex hormones associated with age-related maculopathy in women? The Beaver Dam Eye Study. Trans Am Ophthalmol Soc 1994;92: 289–295. 18. Vaideeswar P, Deshpande JR. Pathology of Takayasu arteritis: a brief review. Ann Pediatr Cardiol 2013;6:52–58. 19. Sugita S, Kawazoe Y, Imai A, et al. Role of IL-22- and TNFalpha-producing Th22 cells in uveitis patients with Behcet’s disease. J Immunol 2013;190:5799–5808. 20. Cruz BA, Reis DD, Araujo CA. Refractory retinal vasculitis due to sarcoidosis successfully treated with infliximab. Rheumatol Int 2007;27:1181–1183. 21. Vila E, Salaices M. Cytokines and vascular reactivity in resistance arteries. Am J Physiol Heart Circ Physiol 2005;288: H1016–H1021. 22. Wagner EM. TNF-alpha induced bronchial vasoconstriction. Am J Physiol Heart Circ Physiol 2000;279:H946–H951. 23. Hao H, Sasongko MB, Wong TY, et al. Does retinal vascular geometry vary with cardiac cycle? Invest Ophthalmol Vis Sci 2012;53:5799–5805. 24. McCanna CD, Myers CE, Lee M, et al. Variability of measurement of retinal vessel diameters. Ophthalmic Epidemiol 2013; 20:392–401. 25. Chandler CS, Gangaputra S, Hubbard LD, et al. Suboptimal image focus broadens retinal vessel caliber measurement. Invest Ophthalmol Vis Sci 2011;52:8558–8561.
