Jugular Venous Reflux Affects Ocular Venous System in Transient Monocular BlindnessChung C.-P.a, c · Hsu H.-Y.c, d · Chao A-C.e · Cheng C.-Y.b · Lin S.-J.c · Hu H.-H.a, c
Departments of aNeurology and bOphthalmology, Taipei Veterans General Hospital,and cInstitute of Clinical Medicine, National Yang-Ming University, Taipei, dSection of Neurology, Department of Internal Medicine, Taichung Veterans General Hospital, Taichung, and eDepartment of Neurology, Kaohsiung Medical University Chung-Ho Memorial Hospital, Kaohsiung, Taiwan Corresponding Author
Background: The frequency of jugular venous reflux (JVR) is higher in patients with transient monocular blindness (TMB). We hypothesize that JVR influences ocular venous outflow, and resulting disturbances in cerebral and ocular venous circulation might be a cause of TMB. To substantiate this hypothesis, we aimed to demonstrate that: (1) TMB patients have vasculature changes in their retinal venules, and (2) JVR could influence ocular venous outflow, as revealed by dilated retinal venules. Methods: This study has 2 parts. The case-control study included 31 TMB patients and 31 age/gender-matched normal individuals, who all received fundus photography for retinal venule diameter comparisons. The Valsalva maneuver (VM) experiment included 30 healthy volunteers who received both color Doppler imaging of the internal jugular vein and fundus photography for retinal venule diameter measurement. Results: In the case-control study, TMB patients had a wider retinal venule diameter (184.5 ± 17.5 vs. 174.3 ± 16.2 µm, right eye, p = 0.023; 194.20 ± 24.6 vs. 176.6 ± 19.5 µm, left eye, p = 0.017), especially TMB patients with JVR. The VM experiments showed that the presence of JVR was associated with a greater increase in retinal venule diameters during VM in the subjects’ right eye (14.27 ± 11.16 vs. 2.75 ± 3.51%, JVR vs. non-JVR, p = 0.0002) and left eye (10.06 ± 6.42 vs. 1.80 ± 2.03%, p = 0.0003). Conclusions: These findings provide evidence that frequently occurring JVR associated with TMB impedes ocular venous outflow, and the subsequent disturbances in ocular venous circulation may be a cause of TMB.
© 2009 S. Karger AG, Basel
Transient monocular blindness (TMB) is defined as sudden, painless, and transient monocular visual loss. The most well-recognized, but not the most common, etiology of TMB is carotid atherothromboembolism [1,2,3,4,5,6]. Therefore, transient ischemic attack should be considered in patients with TMB, and they should be managed promptly [1,2,3,4,5,6]. Although TMB is known to be attributed to ischemia and associated with many conditions, the pathogenesis of TMB remains obscure in up to 66% of these patients [1,2,3,4]. Jugular venous reflux (JVR) indicates an elevated internal jugular vein (IJV) venous pressure, and might impede cerebral venous outflow and retrogradely transmit venous hypertension intracranially [7,8,9,10]. We have found an increased prevalence of JVR in patients with TMB, especially in patients with an undetermined cause and who had had more than 2 TMB attacks . In patients with frequent TMB and attacks of an undetermined cause, the frequency of JVR was 74%, which was higher than the frequency observed (20–40%) in the normal controls [7,11]. Moreover, when they are free from attacks, patients with TMB of an undetermined cause have an increased vascular resistance in the retrobulbar arteries in the absence of significant arterial lesions , which may result from increased venous outflow resistance. Therefore, we hypothesized that JVR would affect ocular venous drainage by impeding venous outflow or transmitting backpressure more easily into the cerebral vessels, and subsequently influencing the ocular venous circulation and lead to TMB. However, the evidence demonstrating a direct causal relationship between JVR and the ocular venous drainage system is lacking.
Human and animal studies examining cerebral venous occlusion have shown that an elevated venous pressure results in a dilatation of upstream venule beds [13,14]. Furthermore, retinal venule dilatation is often associated with ocular venous drainage impairment in certain conditions, such as central retinal vein occlusion or during increased intracranial pressure [15,16]. In the present study, we first compared the retinal venule diameter between TMB patients with an undetermined cause and normal individuals. We hypothesized that TMB patients would have a wider retinal venule compared to age/gender-matched controls, which represents impaired ocular venous drainage in TMB patients. Further, to prove that the JVR frequently found in TMB patients could influence ocular venous outflow, we examined the effect of JVR on the venule diameter changes during the Valsalva maneuver (VM) in normal young individuals in the second part of the study. We hypothesized that increased IJV pressure by VM  would increase retinal venule diameters, and JVR would influence ocular venous outflow as revealed by a more dilated retinal venule diameter during the VM.
TMB Patients versus Age/Gender-Matched Controls. TMB patients were collected from subjects of our previous studies , which prospectively recruited 173 TMB patients from both the outpatients and inpatients of the Neurology Department of Taipei Veterans General Hospital and from referrals for cerebrovascular assessment by ophthalmologists or other physicians. All of these TMB patients received duplex and Doppler ultrasonography of the cervical vessels and transcranial color-coded sonography using a computed sonography system (Acuson; Sequoia, Mountain View, Calif., USA) by the same sonographer. Among these 173 TMB patients, patients who met the following criteria were included in the present study: (1) no identified underlying disease, such as carotid stenosis, heart disease, ophthalmologic diseases, or autoimmune diseases; (2) ≧3 TMB attacks. Forty-six TMB patients with an undetermined cause and frequent attacks were collected in the previous study. Fifteen of these 46 patients were lost to follow-up (10 patients) or were unwilling to undergo further testing (5 patients). Therefore, only 31 patients (15 men, 16 women; mean age: 58.6 ± 17.9 years) were recruited into the present study. These patients received duplex and Doppler ultrasonography of the cervical vessels once more, and an eye fundus photographic examination for retinal venule diameter analysis. All these patients received ophthalmologic examination by ophthalmologists, and all had normal fundoscopic findings and intraocular pressure. For comparison, from people who attended our sonographic laboratory for cervical vessel assessment, we recruited 31 age- and gender-matched normal individuals with normal cervical vessels upon duplex sonographic study (i.e., without carotid stenosis and JVR). Normal individuals who had abnormal visual acuity, an abnormal funduscopic examination, or a history of visual problems were excluded. All individuals received duplex and Doppler ultrasonography of the cervical vessels and transcranial color-coded sonography using the same computed sonography system by the same sonographer. Clinical characteristics of the TMB patients and normal individuals were recorded.
VM Experiment. Thirty healthy young volunteers (14 men, 16 women; mean age: 27.2 ± 3.6 years, range: 20–36 years) were recruited from our hospital staff. None had central nervous system diseases, heart diseases, hypertension, diabetes, hyperlipidemia, malignancy, glaucoma, cataract, or retinal diseases. We did not include older people because they have difficulty performing the VM during fundus photography.
The institutional review board of this hospital approved the study protocol, and written informed consent was obtained from all the participants.
Color Doppler imaging was performed in a head-straight flat supine position with a 7-MHz linear transducer (Acuson) after a 10-min quiet rest for all study participants. Copious amounts of the ultrasound gel were used, and great care was taken to avoid compression of the neck veins during examination. Bilateral IJVs were examined in all individuals at baseline and during VM. At baseline, the IJV was initially insonated with a longitudinal and then a cross-sectional view from the proximal part of the neck base rostrally to the distal part of the submandibular level in order to evaluate the IJV flow pattern by color duplex. Then, we put the distal margin of the window of the color signal at the tip of the flow divider of the internal carotid artery for Doppler spectrum study during VM. The VM was performed for 15 s with a maintained intrathoracic pressure of 40 mm Hg, and was monitored by a pressure gauge connected to a flexible tube. JVR was defined as a characteristic duplex ultrasound and/or Doppler spectrum indicating retrograde flow lasting more than 0.5 s spontaneously (at baseline) or during VM.
TMB Patients versus Age/Gender-Matched Controls. TMB patients and normal individuals all received fundus photography. The fundus photography was performed in a darkened room by a trained physician (C.-P.C.) with a colored fundus camera that does not require pharmacologic dilatation of pupils (Kowa Non Myd 7 digital fundus imaging system; Kowa, Tokyo, Japan). All study individuals had 20-degree color retinal fundus digital photographs taken, centered on the disc at rest for their both eyes. The orders for the right or left eye examinations were random.
VM Experiment. Thirty healthy young volunteers received fundus photography on the same day as IJV color Doppler imaging. All study individuals had 20-degree color retinal fundus digital photographs taken, centered on the disc, at the baseline and during VM for their both eyes. Before the fundus photography, individuals practiced several times to simultaneously perform the VM and look into the objective lens of the fundus camera. Because we had found that a higher VM pressure makes fundus camera alignment difficult in our preliminary experiment, the VM performed here lasted for 10 s with a maintained intrathoracic pressure of only 20 mm Hg, and was monitored by a pressure gauge connected to a flexible tube. Every individual had their fundus photographed both at baseline (at rest) and at the 5–7th s of VM (without leaving the forehead and chin rest of the fundus camera) on one eye, and then went on to the next course for the next eye. The order of these 2 conditions (baseline or VM) for each eye’s photograph and the right or left eye sequences were randomized. If the subject’s first photography session was taken during the VM, the baseline photograph was taken at least 5 min later. All subjects bit the flexible tube, and were asked not to move their heads during the entire course of each eye’s photograph to ensure the same position under these 2 conditions.
All fundus photographs were stored in a computer. Our retinal diagnostic system includes a high-resolution 10-megapixel Nikon D80 camera back and a computer workstation (Kowa Non Myd 7 digital fundus camera and software Kowa VK-2) for image storage and analysis. The digital photograph was taken at 5,908 dots/12.02 mm resolution. The grader overlaid the digital photograph with the standard disc grid (disc diameter: 1,850 µm), then all venules passing through zone B (0.5–1 disc diameter from the disc margin) were measured . The grader chose the segment of each venule within zone B most suitable for measurement (based on image quality and straightness of the vessel) and selected it as the ‘region of interest’. The segment was then magnified 3 times automatically by the computer. The grader determined the vessel edges by locating the points at which the vessel became uniformly darker than the adjacent retinal pigment epithelium and the length between the 2 edges was automatically measured by the computer (accurate to 0.001 mm). A fundus photograph was considered ungradable if more than 1 venule larger than 40 µm in diameter could not be measured due to poor image quality. If an eye, photographed both at baseline and during VM, was considered ungradable, then this eye’s photograph was excluded from the analysis. The Parr-Hubbard formula was used to calculate a standard value of venule diameter in each photograph for analysis . The measurement was performed by a trained grader, who was blinded to the experimental condition for each photograph (baseline or VM), individuals’ jugular venous flow pattern, and the clinical diagnosis. Our previous intra-grader study (n = 20) showed excellent agreement in the retinal venule diameter measurements (intra-class correlation coefficient = 0.92 for the left eye and 0.90 for the right eye).
Continuous data were expressed as means ± SD. For all tests, a value of p < 0.05 was considered statistically significant.
TMB Patients versus Age/Gender-Matched Controls. The comparisons of retinal venule diameter in each eye between TMB patients and normal individuals were analyzed by a Wilcoxon rank-sum test.
VM Experiment. The Wilcoxon signed-rank test was used to evaluate the difference of retinal venule diameters between the baseline and VM in each eye. Because eyes may have a different magnification in cases of refractive changes due to corneal curvature, lens and axial length differences [20,21,22], we used the ratio (venule diameter during VM – venule diameter at baseline)/(venule diameter at baseline) for the retinal venule diameter changes during VM for comparison between eyes. The eye was classified as belonging to the JVR group if JVR had been detected in either side of the IJV. The comparisons of retinal venule diameter changes during VM between right and left eyes, between the JVR group and the non-JVR group, and between genders were analyzed by a Wilcoxon rank-sum test. The correlation analysis between age and the retinal venule diameter changes during VM was calculated using a Pearson correlation analysis.
The clinical characteristics and retinal venule diameter measurements of TMB patients and normal individuals are summarized in table 1. None of the TMB patients and normal individuals had strokes, dementia, malignancies, or heart diseases (valvular diseases, ischemic heart disease, and arrhythmia). All TMB patients experienced an abrupt onset. Twenty patients (64.5%) always had their TMB attacks on the same eye (right eye: 12; left eye: 8), the others had attacks on alternate eyes. Preceding Valsalva-like activities occurred in 10 patients (32.3%), which included exercise (4 patients), bending (5 patients) or straining (8 patients). The duration of the TMB attack was <10 min in 30 patients (96.8%), and >10 min but <1 h in 1 patient. The patterns of visual loss were altitudinal/lateralized in 8 patients (25.8%), diffuse in 20 patients (64.5%), and miscellaneous in 3 patients (9.7%). Accompanied positive visual phenomenon occurred in 5 patients (16.1%). Twenty-two TMB patients (70.97%) had JVR. JVR was detected spontaneously (at baseline) in 10 TMB patients, while the other 12 TMB patients had JVR only during VM. Six TMB patients had bilateral JVR, 10 had JVR only on the left side, and 6 had JVR only on the right side. In TMB patients, there were 2 right-eye and 6 left-eye retinal photographs that were ungradable. In age/gender-matched normal individuals, there were 2 right-eye and 6 left-eye retinal photographs that were ungradable. There was no difference between the right and left retinal venule diameters in both the TMB group (p = 0.098) and control group (p = 0.628). Compared with normal individuals, TMB patients had larger retinal venule diameters at baseline in both the right (184.5 ± 17.5 vs. 174.3 ± 16.2 µm, TMB vs. normal controls, p = 0.023) and left eyes (194.20 ± 24.6 vs. 176.6 ± 19.5 µm, TMB vs. normal controls, p = 0.017). In TMB patients with JVR, more significantly wider retinal venules compared with normal individuals were shown (fig. 1). The retinal venule diameter between symptomatic eye and asymptomatic eye in TMB patients whose TMB attacks were always on the same eye were not different significantly (p = 0.346; fig. 2).
|Table 1. Clinical characteristics and retinal venule diameters|
|Fig. 1. The retinal venule diameters of TMB patients (total), TMB patients (with JVR) and age/gender-matched normal individuals. Left eyes: TMB patients (total; n = 25) 194.2 ± 24.6, TMB patients (JVR; n = 18) 201.9 ± 22.8, normal group (n = 25) 176.6 ± 19.5. Right eyes: TMB patients (total; n = 29) 184.5 ± 17.5, TMB patients (JVR; n = 20) 190.6 ± 13.4, normal group (n = 29) 174.3 ± 16.2.|
|Fig. 2. The retinal venule diameters comparisons between affected eyes and non-affected eyes in TMB patients who always had TMB attacks on the same side (n = 20). There were only 14 individuals with gradable retinal venules in both eyes; 8 had JVR, and among them there were 6 with affected eyes on the right side and 2 on the left side; 6 had no JVR, and among them there were 4 on the right side and 2 on the left side. Affected eyes versus non-affected eyes = 190.6 ± 9.2 versus 195.7 ± 12.6 in the JVR group (n = 8); 178.9 ± 6.8 versus 182.9 ± 13.6 in the non-JVR group (n = 6).|
Baseline characteristics of all individuals are summarized in table 2. Every eye (n = 60) had a gradable fundus photograph at baseline. There was no difference in the retinal venule diameter at baseline between the right and left eyes (175.6 ± 18.7 vs. 180.3 ± 19.4 µm, p = 0.342). During VM, the right eyes had 5 and the left eyes had 9 ungradable fundus photographs due to difficult fundus camera alignment after VM. Retinal venule diameters were increased during VM in most eyes that had gradable fundus photographs both at baseline and during VM, i.e., 20 eyes (80%) on the right and 17 eyes (81%) on the left. There were 5 right eyes and 4 left eyes with unchanged retinal venule diameters during VM. None had decreased retinal venule diameters during VM in either right or left eyes. The median values for retinal venule diameter change during VM (venule diameter during VM – venule diameter at baseline)/(venule diameter at baseline) were 7.47% (range: 0–18%) in the right eye and 4.86% (range: 0–27%) in the left eye. Retinal venule diameter changes during VM were not statistically different between right and left eyes (right vs. left: mean ± SD, 7.36 ± 9.33 vs. 5.73 ± 6.21%; median, 7.47 vs. 4.86%; p = 0.499).
|Table 2. Baseline characteristics of young normal individuals in the VM experiment (n = 30)|
JVR in either one side or both sides of the IJV was detected in 10 individuals (33.3%), and all their JVR were found during VM. Among them, 2 individuals had JVR in both sides of the IJV and the others had JVR in just 1 side of the IJV. There were 8 JVR detected on the right side (8/30) and 4 on the left side (4/30). Retinal venule diameters at baseline were not statistically different between the JVR group and the non-JVR group in both eyes (169.9 ± 24.4 vs. 178.4 ± 15.1 µm, JVR vs. non-JVR, right, p = 0.372; 177.3 ± 18.1 vs. 181.8 ± 20.4 µm, JVR vs. non-JVR, left, p = 0.616). Compared with the non-JVR group, the changes in retinal venule diameter during VM were greater in the JVR group in both the right (JVR vs. non-JVR: mean ± SD, 14.27 ± 11.16% vs. 2.75 ± 3.51%; median 10.54 vs. 1.03%; p = 0.0002) and left eyes (10.06 ± 6.42 vs. 1.80 ± 2.03%; 8.95 vs. 1.24%; p = 0.0003; fig. 3). The retinal venule diameter changes during VM were not associated with gender (right eyes: p = 0.163; left eyes: p = 0.284) or age (right eyes: p = 0.483; left eyes: p = 0.667).
|Fig. 3. Retinal venule diameter changes during VM between the JVR group and non-JVR group in normal young individuals. The bar and the line within the bar represent the range and the median of the retinal venule diameter changes during VM, respectively.|
In the first part of this study, we found that TMB patients with frequent attacks and an undetermined cause had wider retinal venules compared with age- and gender-matched normal individuals. Since venule dilatation often indicates downstream venous outflow impedance [13,14,15,16], the result suggests that TMB patients with frequent attacks and an undetermined cause might have ocular venous drainage impairment. Furthermore, more significantly wider retinal venules were found in TMB patients with JVR (fig. 1), these results support our hypothesis that JVR plays a major role in the pathogenesis of TMB by impeding ocular venous outflow or transmitting backpressure more easily into the cerebral vessels.
To prove that JVR does have an influence on retinal venule diameter, we conducted the second part of the present study. In the VM experiment, we found that: (1) retinal venule diameter increased during VM, and (2) JVR produced a greater increment in retinal venule dilatation during VM. These findings suggest that IJV venous hypertension induced by VM could affect ocular venous drainage. Furthermore, our results showed that JVR produced a greater retinal venule dilatation during VM, which clearly provides evidence indicating that JVR affects ocular venous drainage. For paired analysis of retinal venule diameters between the baseline and VM photographs, one concern included the different levels of magnification between these 2 conditions in each eye. We minimized this magnification error by asking study subjects to not move their heads during the entire course of each eye’s photographic session so as to ensure the same eye position under these 2 conditions. Furthermore, we found that the vertical optic disc diameters were similar between these 2 conditions in both the right and left eyes (1.73 ± 0.22 vs. 1.74 ± 0.23 mm, baseline vs. VM, right, p = 0.354; 1.70 ± 0.20 vs. 1.70 ± 0.20 mm, baseline vs. VM, left, p = 0.898). Therefore, the observed increase in retinal venule diameter after VM should not be biased by magnification errors.
Unlike cerebral arterial disease, which usually causes cerebral dysfunction ipsilateral to 1 stenotic or occluded artery, cerebral venous outflow impedance, even in 1 side, would induce global cerebral venous hypertension . Therefore, 1-sided JVR in our TMB patients or normal subjects in VM experiments could lead to contralateral eye TMB attacks  and bilateral retinal venule dilatation. However, there were 20 of our TMB patients who had always TMB attacks in the same eye, and their retinal venule diameter was similar to the other eye without TMB attack. Whether or not the affected eye was the dominant eye or had lower tolerance to venous ischemia needs further study.
The VM experiment demonstrated that normal young people with JVR did not have wider retinal venules at baseline, but only during VM. In contrast, our TMB patients already had wider retinal venules at baseline when compared with normal individuals. This might be due to the fact that the TMB patients were older and the JVR might have pre-existed for longer periods of time. Moreover, JVR was detected in 10 (32.26%) of our TMB patients at baseline, who would have been exposed to longer periods of retrograde-transmitted venous pressure via IJV than people with JVR only during VM. Another explanation is that JVR in TMB patients might produce higher retrograde-transmitted venous pressures and more severe cerebral venous outflow impairment than in normal young individuals. However, this line of investigation needs further quantification in order to be confirmed.
Retinal venule dilatation has been associated with diabetes , renal disease , atherosclerosis , cerebral small vessel disease , incident strokes , obesity , and hypertriglyceridemia . There were no differences in the frequencies of these vascular risk factors between TMB and normal groups (table 1). Besides, we presume that the bias of retinal venule diameter measurements due to differences of corneal curvature, lens, and axial length were distributed equally across the TMB and normal groups. However, the magnification errors still may be a concern.
There are limitations to our study. Blood supply to the eye depends on the retrobulbar artery pressure, arterial flow and the venous flow, which we did not measure in the present study. However, all our patients had normal carotid artery flow and arterial blood pressure, which leads to the reasonable assumption that the retinal artery blood flow and pressure were similar with the normal individuals. As to the ophthalmic vein, we did check the venous flow in our previous study, which showed a higher frequency of reversal of superior ophthalmic venous flow in undetermined cause/high frequency TMB patients .
Interestingly, vasospasms have been shown to be a cause of TMB [1,31,32]. It has been suggested that disturbances in venous outflow might precipitate vasospasms [11,33], thereby indicating that vasospasms may be a consequence rather than a causal event . Further studies analyzing the association between the venous flow patterns and changes in the retinal arteriole diameter are required to test this hypothesis.
In conclusion, JVR was frequently found in TMB patients of an undetermined cause with frequent attacks. Our TMB patients with an undetermined cause and frequent attacks had larger diameter retinal venules, especially patients with JVR. We have provided evidence demonstrating that JVR affects ocular venous drainage and may contribute to dilated retinal venules in the long run. These results support our previous hypothesis that frequent JVR found in TMB impedes ocular venous outflow, and disturbances of cerebral and ocular venous circulation might be an underlying cause of TMB, especially in patients who suffer from frequent attacks with an undetermined cause.
This work was supported by the National Science Council Research (NSC95-2314-B075-041-MY3, H.-H.H.).
Prof. Han-Hwa Hu
Department of Neurology, Taipei Veterans General Hospital
201 Sec. 2, Shihpai Road
Peitou, Taipei 11217 (Taiwan)
Tel. +886 2 2875 7046, Fax +886 2 2873 9241, E-Mail email@example.com
Received: April 6, 2009
Accepted: August 8, 2009
Published online: December 1, 2009
Number of Print Pages : 8
Number of Figures : 3, Number of Tables : 2, Number of References : 33
Vol. 29, No. 2, Year 2010 (Cover Date: January 2010)
Journal Editor: Hennerici M.G. (Mannheim)
ISSN: 1015-9770 (Print), eISSN: 1421-9786 (Online)
For additional information: http://www.karger.com/CED