Color fundus photography and 3 × 3-mm en face optical coherence tomography angiography images illustrating a papillomacular (A), inferotemporal (B), and superotemporal (C) RVM. The RVM can bypass or cross the fovea without affecting visual acuity.
Fluorescein angiography (FA) frames illustrating the venous nature and course of the retinal malformation in various cases. A and B, Early and late FA frames illustrating venous filling in the retinal venous malformation. C and D, Fluorescein angiogram illustrating the venous laminar pattern of filling in the congenital retinal macrovessel during the transit phase. Note the various courses of the retinal macrovessels, which may cross the fovea (A), bypass the fovea (B), or encircle the fovea (C and D). As can be seen, congenital retinal macrovessels are venous malformations of the retina.
Color fundus photography and 3 × 3-mm en face optical coherence tomography angiography segmented at the level of the deep retinal capillary plexus in 2 patients with retinal venous malformations. Note the presence of microvascular dilation that is more readily identified in the deep retinal capillary plexus vs the superficial retinal capillary plexus shown in Figure 1. Optical coherence tomography angiography of retinal venous malformations identifies abnormalities of the deep capillary plexus.
Fluorescein angiography (A), spectral-domain optical coherence tomography (OCT) (B), and 3 × 3-mm en face OCT angiography (C) of a patient with subinternal limiting membrane hemorrhage in an eye with retinal venous malformation. A, Early fluorescein angiography illustrates a superior retinal venous malformation associated with an abnormal dilated microvascular plexus. B, The level of spectral-domain OCT scans is indicated by the colored lines in A. Note the subinner limiting membrane hemorrhage illustrated in the bottom panel and the dilated deep capillary plexus illustrated in the top panel. C, 3 × 3-mm En face OCT angiogram of the deep retinal capillary plexus illustrates a dilated microvascular plexus at the level of the deep capillary plexus.
Magnetic resonance imaging (left) and color fundus photography (right) of 4 patients with retinal venous malformation that illustrate coexistent retinal venous malformations of the brain. Arrowheads indicate venous malformations in the brain parenchyma; circles indicate cavernomas.
eTable 1. Ophthalmic complications in 9 eyes with retinal venous malformations.
eTable 2. Current classifications of vascular malformations of the brain.
eFigure 1. Color fundus photography and en face optical coherence tomography angiography of retinal venous malformations.
eFigure 2. Color fundus photography, fluorescein angiography, and spectral-domain optical coherence tomography of central serous chorioretinopathy in an eye with retinal venous malformation.
eFigure 3. Incidental finding of a retinal venous malformation in a patient with acute macular neuroretinopathy.
eFigure 4. Magnetic resonance imaging and color fundus photography of a case of retinal venous malformation.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Pichi F, Freund KB, Ciardella A, et al. Congenital Retinal Macrovessel and the Association of Retinal Venous Malformations With Venous Malformations of the Brain. JAMA Ophthalmol. 2018;136(4):372–379. doi:10.1001/jamaophthalmol.2018.0150
What is the association of congenital retinal macrovessel with venous anomalies of the brain?
In this cross-sectional series of 49 patients with congenital retinal macrovessel, 12 patients (24%) were noted to have associated vascular (typically venous in nature) malformations of the brain on magnetic resonance imaging. Further, congenital retinal macrovessels were of venous origin.
These findings suggest imaging studies of the brain should be considered in patients with congenital retinal macrovessel, which may more accurately be referred to as retinal venous malformations.
Congenital retinal macrovessel (CRM) is a rarely reported venous malformation of the retina that is associated with venous anomalies of the brain.
To study the multimodal imaging findings of a series of eyes with congenital retinal macrovessel and describe the systemic associations.
Design, Setting, and Participants
In this cross-sectional multicenter study, medical records were retrospectively reviewed from 7 different retina clinics worldwide over a 10-year period (2007-2017). Patients with CRM, defined as an abnormal, large, macular vessel with a vascular distribution above and below the horizontal raphe, were identified. Data were analyzed from December 2016 to August 2017.
Main Outcomes and Measures
Clinical information and multimodal retinal imaging findings were collected and studied. Pertinent systemic information, including brain magnetic resonance imaging findings, was also noted if available.
Of the 49 included patients, 32 (65%) were female, and the mean (SD) age at onset was 44.0 (20.9) years. A total of 49 eyes from 49 patients were studied. Macrovessel was unilateral in all patients. Color fundus photography illustrated a large aberrant dilated and tortuous retinal vein in all patients. Early-phase frames of fluorescein angiography further confirmed the venous nature of the macrovessel in 40 of 40 eyes. Optical coherence tomography angiography, available in 17 eyes (35%), displayed microvascular capillary abnormalities around the CRM, which were more evident in the deep capillary plexus. Of the 49 patients with CRM, 39 (80%) did not illustrate any evidence of ophthalmic complications. Ten patients (20%) presented with retinal complications, typically an incidental association with CRM. Twelve patients (24%) were noted to have venous malformations of the brain with associated magnetic resonance imaging. Of these, location of the venous anomaly in the brain was ipsilateral to the CRM in 10 patients (83%) and contralateral in 2 patients (17%), mainly located in the frontal lobe in 9 patients (75%).
Conclusions and Relevance
Our study has identified an association between macrovessels in the retina and venous anomalies of the brain (24% compared with 0.2% to 6.0% in the normal population). Thus, we recommend new guidelines for the systemic workup of patients with CRM to include brain magnetic resonance imaging with contrast. These lesions may be more accurately referred to as retinal venous malformations, which may raise awareness regarding potential cerebral associations.
In 1869, Mauthner1 was the first to describe a large aberrant retinal vessel in the macula. However, Brown et al2 later coined the term congenital retinal macrovessel (CRM) in a report describing 7 cases of an abnormal retinal vessel, typically a vein, crossing the central macula with a vascular distribution above and below the horizontal raphe. Archer et al3 subsequently classified CRM as a form (stage 1) of arteriovenous malformation of the retina.
Since the first description, to our knowledge, fewer than 40 cases4,5 of CRM have been reported. The prevalence of CRM is rare and has been estimated at approximately 1 in 200 000 in the United States,4 but this may be an underestimation of the true number because CRMs are typically asymptomatic and diagnosed during routine retinal examination. Vision loss is rarely encountered and may be attributed to macular hemorrhage,6 serous macular detachment,7,8 or foveal cystoid formation.9 The aberrant retinal vessel may cause vision loss if it crosses the foveal region.
To our knowledge, multimodal imaging has not been studied in the evaluation and management of CRM; the diagnosis can be readily made with clinical examination when the macrovessel transverses the horizontal raphe. Traditional consensus is that lesions are isolated and do not require systemic workup, such as brain magnetic resonance imaging (MRI). Because of the paucity of reports in the literature and the recent advancement in retinal imaging, we aimed to study a large series of eyes with CRM with advanced multimodal retinal and systemic imaging, including brain MRI, to better understand the pathogenesis and clinical spectrum associated with this disorder.
As a result of the findings in this article, including the retinal venous origin of all lesions and the high association with venous abnormalities of the brain, we recommend referring to CRM as retinal venous malformations (RVM), which is the nomenclature we will use throughout the article going forward.
This was an institutional review board–approved retrospective review of patients with clinical evidence of an RVM. This study was approved by the Cleveland Clinic Abu Dhabi Institutional Review Board, complied with the Health Insurance Portability and Accountability Act of 1996, and followed the tenets of the Declaration of Helsinki. Because of the retrospective nature of this study, informed consent was waived.
Diagnosis of RVM was based on the characteristic clinical findings—namely, an abnormal, large, retinal vessel crossing the macular region with a vascular distribution above and below the horizontal raphe. Exclusion criteria included poor-quality imaging precluding optimal analysis of retinal findings or the presence of other retinal vascular disorders, such as diabetic retinopathy, or retinal vein occlusion not directly related to the underlying disorder of RVM.
Patients with RVM referred to 7 different retina clinics worldwide over the 10-year period between 2007 and 2017 were identified from local databases, and clinical information and multimodal retinal imaging findings were collected and studied. Medical records were reviewed and analyzed to identify the nature, predisposing factors, and course of the lesions. Pertinent systemic information, including brain MRI findings, was also noted if available.
Demographic information, including age and sex at baseline presentation, was collected. Ophthalmic findings at the time of presentation, including best-corrected visual acuity and findings with slitlamp and dilated retinal examination, were recorded. When available, the findings of follow-up examinations were recorded to determine the natural course of uncomplicated RVM and the incidence, nature, and evolution of complications when present.
Color fundus photography was obtained with CIRRUS Photo (Carl Zeiss Meditec), a conventional 9-field fundus camera, or with the Optos Panoramic 200MA (Optos PLC), a wide-field fundus photography system. Fluorescein angiography (FA) was performed with either VISUCAM (Carl Zeiss Meditec) or Spectralis HRA+OCT (Heidelberg Engineering), 9-field FAs, or with Optos Panoramic 200MA, a wide-field system. Indocyanine green angiography of the posterior pole and the 9 peripheral fields was acquired with the Spectralis HRA+OCT. Fundus autofluorescence images were obtained with the Spectralis HRA+OCT or the Optos Panoramic 200MA. A custom cross-line and a 20° × 20° volume acquisition protocol were used to obtain a set of high-speed scans from each eye with spectral-domain optical coherence tomography (OCT) imaging (Spectralis HRA+OCT and Cirrus HD-OCT 4000 version 5.0 [Carl Zeiss Meditec]).
The Optovue RTVue XR Avanti (Optuvue Inc) and the Zeiss AngioPLEX (Carl Zeiss Meditec) were used to acquire en face OCT angiography images of the RVM. A 3 × 3-mm and/or a 6 × 6-mm scanning area centered on the fovea was captured for blood flow measurements and standardized segmentation of the superficial and deep retinal capillary plexus in each patient.
A total of 49 eyes from 49 patients were included in this study. Seventeen patients (35%) were male and 32 (65%) were female. Six (12%) were Hispanic, 7 (14%) were Arabic, 2 (4%) were Asian, 10 (20%) were African American, and 24 (49%) were white. The mean (SD; range) age at onset was 44.0 (20.9; 18.0-85.0) years. Retinal venous malformation was unilateral in all patients and involved the right eye in 15 patients (31%) and the left eye in 34 patients (69%). Visual acuity at presentation ranged from 20/20 OU to 20/400 OU (median, 20/25 OU).
Of the 49 patients RVM, 39 (80%) did not have any evidence of ophthalmic complications. The mean best-corrected visual acuity in these 39 eyes was 20/25 OU, and within this group, there was no statistically significant difference in best-corrected visual acuity in eyes with RVM crossing vs not crossing the fovea.
Color fundus photography was available for 45 patients (92%; 36 [80%] 9-field and 9 [20%] wide-field fundus photography) and illustrated a large aberrant dilated and tortuous retinal vein in all patients, located either inferotemporal (27 of 49 [55%]) or superotemporal (17 of 49 [35%]) in the macular region or in the papillomacular bundle (5 of 49 [10%]) (Figure 1). The RVM branched inferiorly or superiorly and crossed the horizontal raphe in all patients. Multiple tributaries were identified in the macula; these tributaries bypassed the fovea in 7 patients (14%), encircled the fovea in 13 patients (27%), and crossed through the fovea in 29 patients (59%) (Figure 1).
Fluorescein angiography was available for 40 eyes (82%); of these, 36 patients (90%) had baseline 9-field FA data (9 on VISUCAM and 27 on Spectralis HRA+OCT), whereas 4 (10%) had wide-field FA data (Optos Panoramic 200MA). Early frames of FA (Figure 2) further confirmed the venous nature of the RVM in these 40 eyes. Fluorescein angiography of RVM without retinal complications displayed early filling and late staining of the aberrant vessel in all patients and illustrated dilation of an anomalous capillary network associated with the RVM in all of the uncomplicated eyes without evidence of leakage (Figure 2). A few microaneurysms were detected in the capillary network around the RVM in 3 of the 39 uncomplicated eyes (8%). Leakage from the RVM and associated macular ischemia were notably absent in all patients.
Spectral-domain OCT scans were available in all patients with sections through the RVM in 10 patients (20%). The main trunk of the RVM was located in the inner plexiform layer and produced significant signal attenuation or shadowing in the structures below. The minor venular branches could be localized by spectral-domain OCT in the inner plexiform layer in 29 patients (59%) and in the inner nuclear layer in 20 patients (41%).
This was confirmed with OCT angiography analysis, available in 17 eyes (35%). Optical coherence tomography angiography displayed an increase in the vascular flow in RVM compared with the surrounding normal veins. The larger macrovessel was clearly identified in the superficial capillary plexus slab (Figure 1) in 13 eyes and illustrated microvascular capillary abnormalities around the RVM, which were more evident in the deep capillary plexus (DCP) compared with the superficial capillary plexus slab (Figure 3; eFigure 1 in the Supplement). Indocyanine green angiography (8 of 49 eyes [16%]) and fundus autofluorescence (24 eyes [49%]) were noncontributory.
eTable 1 in the Supplement reports the retinal complications affecting 10 of 49 eyes (20%), the demographic information of these patients, and their treatment. The mean best-corrected visual acuity of these 10 eyes was 20/40 OU (Figure 4).
Two patients (men in their late 30s and early 40s, respectively) with RVM developed acute central serous chorioretinopathy. Neither patient reported any recent stressful event or corticosteroid exposure. Best-corrected visual acuity was 20/40 OD and 20/30 OS in both patients. Dilated retinal examination displayed a serous macular detachment and associated retinal pigment epithelium changes at the margin of the fluid in each patient (eFigure 2 in the Supplement). Fluorescein angiography confirmed the diagnosis of central serous chorioretinopathy associated with an RVM, and mottled hyperfluorescence was present in the area corresponding to the pigmentary changes in each patient. There was no demonstrable leakage or arteriovenous anastomosis associated with the RVM. With spectral-domain OCT, a serous retinal detachment was detected in both patients, and an associated pigment epithelial detachment was noted in patient 1. On enhanced depth imaging OCT, the choroid underlying the detachment was thickened (eFigure 2 in the Supplement) in both patients.
Patient 4, a young woman, presented with Dengue fever and a dense central scotoma of the right eye. Best-corrected visual acuity was 20/40 OD. Fundus photography illustrated a venous RVM arising from the inferior branch of the central retinal vein with capillary branching and anastomosis in the perifoveal area (eFigure 3 in the Supplement). Spectral-domain OCT illustrated nasal disruption of the ellipsoid layer (eFigure 3 in the Supplement), and irreversible thinning of the outer nuclear layer ensued, consistent with acute macular neuroretinopathy. After 1 week of oral prednisone, 1 mg/kg, the external limiting membrane and ellipsoid zone were improved, but persistent disruption was noted.
Two patients (woman and man, respectively, in their 50s) presented to the ophthalmologist complaining of vision loss (20/200 OU and 20/400 OU) due to intraretinal hemorrhage in the foveal area associated with a venous RVM. Fundus photography illustrated RVM complicated by hemorrhage and an associated saccular dilation in both patients at the level of the deep retinal capillary plexus. Fluorescein angiography localized the aneurysm underneath the RVM in patient 8 (Figure 1A; Figure 2A), and spectral-domain OCT illustrated the aneurysm at the level of the inner nuclear layer in each patient. Both patients were observed with spontaneous involution of the DCP aneurysm and reabsorption of the blood at a mean of 4.5 months.
Of the 49 patients included in this series, 13 (27%) presented with a history of hypertension, 4 (8%) with a history of diabetes, and 8 (16%) with a history of hyperlipidemia. Per exclusion criteria, none of the 17 patients with diabetes and hypertension illustrated any retinal manifestations of these systemic diseases. Two patients from Brazil presented with acute macular neuroretinopathy and Dengue fever; association with RVM was incidentally diagnosed. One patient presented with acute myelogenous leukemia.
Previous brain MRI was available for 27 patients (55%) in the present case series. In 25 of 27 patients (93%), the imaging studies of the brain were ordered by a general practitioner, a neurologist, or an internist, independent of the retinal findings. The main reason for requesting brain MRI was long-term headache in 18 patients (67%), followed by otological indications in 4 (15%), endocrinology screening of the pituitary gland in 2 (7%), behavioral changes in 1 (4%), suspicion of idiopathic intracranial hypertension in 1 (4%), and benign tumor of the brain in 1 (4%). In 2 patients (7%), the retina specialist requested brain imaging after the diagnosis of RVM.
Of the 27 brain MRIs that were reviewed, 10 (37%) were considered unremarkable by the reading neuroradiologist. Five patients (19%) had findings consistent with the underlying disease for which the MRI was obtained (1 with ear abnormalities, 1 with pituitary abnormalities, 2 with old ischemic lesions, and 1 with a benign brain tumor).
Twelve patients (24%) were noted to have associated neurovascular lesions of the brain (Figure 5; eFigure 4 in the Supplement). Brain MRI with contrast was performed in all of these patients, and additional brain angiograms were requested in 4 (33%). “Venous malformation” was diagnosed in 6 patients (50%), “arteriovenous malformation” in 2 (17%), and “developmental venous malformation associated with isolated cavernoma” in 4 (33%). In all patients, the venous anomaly was detected in the T1 postcontrast MRI images, while the 4 cavernomas were detected in T2 images (Figure 5). Location of the venous anomaly in the brain was ipsilateral to the RVM in 10 patients (83%) and contralateral in 2 patients (17%), mainly located in the frontal lobe in 9 patients (75%). The mean size of the cavernomas was 0.7 cm, and they were mainly centered in the anterior cingulate gyrus.
With 49 patients with RVM identified in 15 retina clinics over 4 continents, to our knowledge, our study is the largest study of RVM reported in the literature. A PubMed search using the keyword retinal macrovessel produced 38 articles,6-16 mostly case reports, for a total of 43 patients with RVM reported. A prior series4 from 2008 reported 13 patients with RVM, all originating from a retinal vein. In a retrospective analysis evaluating fundus photographs of 3506 eyes,4 7 eyes in 6 patients (4 men and 2 women) illustrated RVM only as a branch of the inferotemporal arcade supplying the superior retina. Indeed, in our series, there was a prevalence in the inferotemporal distribution of RVM compared with the superotemporal quadrant (60% vs 33%).
Retinal venous malformations are typically benign and are often detected during routine examination.17 Of the 43 patients reported in the literature, 22 (51%) were associated with ophthalmic complications. The 39 uncomplicated cases (80%) in the present series were all detected incidentally during routine retinal examination and proved to be stable, with no complications during follow-up. In our series, 10 patients (20%) presented with retinal complications. Most of these complications were likely an incidental association with RVM.
An RVM is in fact usually a retinal vein and rarely an artery. Of the 38 articles in the literature, only 4 RVMs originating from a retinal artery are reported.18-24 However, none of these 4 cases included an FA study to confirm the true arterial nature. Fluorescein angiography was available in 40 eyes (82%) in our series, and the venous origin of the RVM was confirmed in all cases (Figure 2). The 9 remaining patients did not undergo FA; however, fundus photography of these 9 patients were reviewed by 2 trained retina specialists (F.P. and P.N.), and an agreement of 100% was noted regarding the venous origin of the RVM. In our case series, we could not identify any RVM originating from an artery.
Our series also studied OCT angiography changes associated with the RVM (17 eyes [35%]).25 While the RVM was present at the level of the superficial capillary plexus slab and DCP, likely related to the course of the RVM through the retina, dilation and microvascular abnormalities were predominantly identified in the DCP, which confirms the venous nature of these aberrant vessels (Figure 3) and is consistent with the current theory that the DCP is predominantly responsible for venous drainage.20
The most important finding in our series was the detection of vascular malformations of the brain in 12 patients with RVM (24%). Vascular malformation of the brain is an umbrella term that includes at least 6 neurovascular disorders.26 Such malformations are classified into several types in which the symptom, severity, and cause may vary (eTable 2 in the Supplement).21 Historically, RVMs have never been associated with systemic vascular abnormalities, and to our knowledge, there are no current recommendations to obtain brain MRI on detection of an RVM. In 2015, Sanfilippo and Sarraf15 described a case of RVM associated with ipsilateral brain venous malformation. That patient was included in the current case series. Our study, for the first time to our knowledge, has identified a strong association of RVM with vascular (especially venous) malformations of the brain, and approximately 1 of 4 patients were so affected in our series. Retinal venous malformations seem to be associated in particular with venous anomalies (24%) that are part of the spectrum of arteriovenous malformations. These neurovascular lesions may be accompanied by cavernous malformations (also called cavernous angiomas, cavernous hemangiomas, or cavernomas),27,28 abnormally enlarged collections of blood-filled spaces, as we reported in 4 of 49 patients (8%) in our series. Our study validates this important association between macrovessels in the retina and venous anomalies of the brain, which was first reported by Sanfilippo and Sarraf.15 Thus, we recommend new guidelines for the workup of patients with RVM to include brain MRI with contrast to rule out venous malformations of the central nervous system.
Further, the term macrovessel seems to be a misnomer for this retinal entity. The close association with venous malformations of the brain and the consistent venous nature of the anomalous retinal vessels indicate that retinal venous malformation may be a much more appropriate name and may raise awareness regarding potential cerebral associations.29-31
Our study had limitations, including the retrospective and uncontrolled methods. Most of the patients were ascertained in tertiary care centers, which may represent a referral bias. A larger general population study may provide a more accurate assessment of the true incidence of coexisting retinal and brain venous malformations.
This study validates the importance of a change in nomenclature for this entity. We recommend that congenital retinal macrovessels henceforth be referred to as retinal venous malformations, a more precise and accurate terminology that highlights the venous nature of these lesions and the potential cerebral associations. With the benefit of multimodal retinal imaging, these unilateral developmental lesions were venous in all 49 patients. One-quarter of patients (12 of 49) with RVM harbored a significant systemic vascular abnormality on brain MRI, and a venous malformation was noted in 10 patients. The reported rate of brain venous malformations in the normal population varies from 0.2% to 6.0%, depending on the series. It is well established that the ophthalmologist may play a critical role in the accurate and early diagnosis of systemic (central nervous system and cutaneous) vascular anomalies associated with phakomatosis syndromes and tumors of the retina.32 With the findings reported in this series, we propose that every case of RVM should prompt an MRI study of the brain.
Accepted for Publication: January 9, 2018.
Published Online: March 1, 2018. doi:10.1001/jamaophthalmol.2018.0150
Correction: This article was corrected on September 6, 2018, to include conflicts of interest that were disclosed but inadvertently omitted during the editing process.
Corresponding Author: Francesco Pichi, MD, Cleveland Clinic Abu Dhabi, Eye Institute, PO Box 112412, Abu Dhabi, United Arab Emirates (firstname.lastname@example.org).
Author Contributions: Drs Pichi and Sarraf had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Pichi, Freund, Ciardella, Dackiw, Arevalo, Villarreal, Neri.
Acquisition, analysis, or interpretation of data: Pichi, Freund, Ciardella, Morara, Abboud, Ghazi, Dackiw, Choudhry, Cunha de Souza, Provetti Cunha, Liu, Wenick, He, Villarreal, Sarraf.
Drafting of the manuscript: Pichi, Ciardella, Morara, Ghazi.
Critical revision of the manuscript for important intellectual content: Pichi, Freund, Ciardella, Abboud, Ghazi, Dackiw, Choudhry, Cunha de Souza, Provetti Cunha, Arevalo, Liu, Wenick, He, Villarreal, Neri, Sarraf.
Statistical analysis: Pichi.
Administrative, technical, or material support: Morara, Abboud, Dackiw, Choudhry, Cunha de Souza, Provetti Cunha, Liu, He.
Study supervision: Freund, Ciardella, Ghazi, Arevalo, Neri, Sarraf.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Freund has received personal fees from Heidelberg Engineering, Optos, and Optovue during the conduct of the study as well as from Spark Therapeutics and Genentech/Roche. Dr Choudhry has received nonfinancial research support from Topcon Medical Systems. Dr Arevalo has received nonfinancial support from EY Engineering; personal fees for consulting from Turing Pharmaceuticals, Second Sight Medical Products, Dutch Ophthalmic Research Center International, Allergan, and Bayer; and royalties from Springer. Dr Sarraf has received consultation fees from Amgen, Bayer, Genentech, Novartis, Nuvelution Pharmaceuticals, and Optovue; grants from Allergan, Genentech, Optovue, and Regeneron; and nonfinancial research support from Heidelberg Engineering and Optovue. No other disclosures were reported.
Create a personal account or sign in to: