Mean Visual Function Questionnaire scores. CVO indicates central retinal vein occlusion.
More mean Visual Function Questionnaire scores. CVO indicates central retinal vein occlusion.
Deramo VA, Cox TA, Syed AB, Lee PP, Fekrat S. Vision-Related Quality of Life in People With Central Retinal Vein Occlusion Using the 25-Item National Eye Institute Visual Function Questionnaire. Arch Ophthalmol. 2003;121(9):1297-1302. doi:10.1001/archopht.121.9.1297
To study visual function and vision-related quality of life in persons with central retinal vein occlusion using the 25-item National Eye Institute Visual Function Questionnaire (VFQ-25).
Interviewer-administered study of persons with central retinal vein occlusion. Scores on the VFQ-25 were analyzed and converted to a 100-point scale in which 100 represents the best possible score and 0 represents the worst. Subscale results were compared with previously published data, and a subgroup analysis was performed.
Fifty-one patients participated. The mean adjusted subscale responses were significantly lower than those from a reference group of patients without ocular disease but not significantly different from patients with diabetic retinopathy. Responses to the VFQ-25 correlated most strongly with visual acuity in the better-seeing eye, number of systemic medical conditions, and patients' opinions about their general health. Responses did not correlate with visual acuity in the involved eye.
Central retinal vein occlusion is an ocular disease that is associated with a decreased vision-related quality of life as measured by the VFQ-25. A decrease in VFQ-25 scores is related to the degree of visual loss in the better-seeing eye and the overall systemic health of the patient.
RETINAL VEIN occlusion is the second most common retinal vascular cause of visual loss after diabetic retinopathy.1 Central retinal vein occlusion (CVO) typically affects relatively older persons; however, it may occur in younger individuals. It can lead to severe visual loss by causing macular edema, macular ischemia, vitreous hemorrhage, or neovascular glaucoma. Although most individuals have unilateral disease, the risk of bilateral involvement is approximately 1% per year.2
The National Eye Institute Visual Function Questionnaire (VFQ) was developed to test the psychometric properties of diseases that cause vision loss. The survey was designed to evaluate vision-related quality of life (QOL). Reliable and valid questionnaires with 51 questions (VFQ-51)3 and 25 questions (VFQ-25)4 have been created and used to survey patients with cataract, age-related macular degeneration, diabetic retinopathy, primary open-angle glaucoma, cytomegalovirus retinitis, low vision, optic neuritis, uveitis, and a reference group of individuals without eye disease.3- 9 We studied visual function and QOL in patients with CVO using the VFQ-25.
All patients with a clinical diagnosis of CVO (International Classification of Diseases, Ninth Revision 362.35) were identified by a search of the computerized patient database of the Duke University Eye Center (Durham, NC) from August 1998 through July 1999. The single inclusion criterion was the presence of CVO in at least one eye. Patients were excluded from study entry if they were younger than 18 years or there was no recorded examination by a retinal specialist at our institution to verify the diagnosis.
After approval from the institutional review board, patients were asked to participate, either during their routine clinic appointment or by telephone interview, in a voluntary 20-minute survey. One of the authors (V.A.D. or A.B.S.) administered the survey. The interviewer was not the patient's regular physician.
The VFQ-25 addresses 12 subscales: General Health (1 question), General Vision (1 question), Near Vision (3 questions), Distance Vision (3 questions), Driving (2 questions), Peripheral Vision (1 question), Color Vision (1 question), Ocular Pain (2 questions), Role Limitations (2 questions), Dependency (3 questions), Social Function (2 questions), and Mental Health (4 questions).4 Answers to each question on the VFQ-25 were converted to a 100-point scale in which 100 represents the best possible score and 0 represents the worst. Each subscale represents the average of 1 or more questions, as previously described.3,4 Patients were also asked whether they were currently employed and whether they lived alone. Clinical variables, including patient age, sex, visual acuity, laterality, presence of ischemic CVO, duration of the disorder, and number of systemic medical conditions, were obtained from the medical records.
To correlate visual acuity with subscale responses, the Snellen visual acuity in the affected eye(s) was converted to a log MAR equivalent. Approximations for visual acuity worse than 20/400 were as follows: counting fingers, 20/2000; hand motions, 20/4000; light perception, 20/8000; and no light perception, 20/16000.10 If a patient developed bilateral involvement after participating in the study, the VFQ-25 was repeated. In those cases, the later (bilateral) responses were used in the overall data group. Unilateral responses were analyzed separately.
Patient data from this study was compared with previously known data from a reference group without ocular disease, persons with diabetic retinopathy, and individuals with low vision from a variety of causes.4 Statistical analysis was performed using SAS 8.0 software (SAS Institute, Cary, NC). The t test was used to compare subscale results and patient age in patients with CVO with those in patients in the comparison groups. Because of the small number of patients, statistical comparisons were not performed for patients with bilateral CVO. Univariate analysis using the Wilcoxon rank sum test was performed to determine if there was a significant difference between subscale responses in patients with CVO based on sex, presence of ischemic CVO, current employment, or living alone. No corrections were made for multiple comparisons. Association between subscale responses and other variables (visual acuity, number of systemic medical conditions, age, and duration of vision loss) was assessed using the Kendall tau-b correlation coefficient. Associations involving controlled variables were assessed using the Spearman correlation.
Fifty-one consecutive patients participated in this study. Forty-eight patients had unilateral CVO, and 5 had bilateral CVO. Two patients were surveyed twice, initially with unilateral CVO and again after developing bilateral CVO. Data on patient characteristics are presented in Table 1. No patient declined to be interviewed.
Summaries of the mean adjusted subscale responses are presented in Table 2. Patient responses were compared with previously published data.4 Overall, scores were significantly lower in CVO patients than in a reference group without ocular disease (all except Ocular Pain) but were not statistically different from patients with diabetic retinopathy (except Peripheral Vision and Color Vision) (Figure 1). Scores were lower in individuals with bilateral CVO than in those with unilateral disease, and were generally similar to known means in patients with low vision (Figure 2). Comparison of patient age in the different series found that persons with CVO were significantly older than those with diabetic retinopathy and reference patients (mean age, 69.5, 57, and 59 years, respectively; P<.001 for each comparison).
Univariate analysis using the Wilcoxon rank sum test was performed to determine if ischemic CVO, sex, whether the patient lived alone, and employment status affected VFQ-25 subscale responses (Table 3). There were significantly lower responses for the Dependency subscale in patients who lived alone (P = .05), for Distance Vision in patients who were not employed (P =.02), and for Ocular Pain in patients who were employed (P = .02). The data suggested a lower response for General Health and Driving in patients who were not employed. Other comparisons between the groups were not significantly different.
A comparison of visual acuity and subscale responses (Table 4) indicated that no correlation was seen when visual acuity in the worse eye was considered in all 51 patients. However, a strong correlation was seen for 9 of the 12 subscales when visual acuity in the better eye was considered. In patients with unilateral CVO, the worse-seeing eye was the one with CVO in 45 of the 46 patients.
To determine the effect of unilateral CVO on vision-related QOL, and to avoid the confounding effect of poor vision in the unaffected eye, we performed a subgroup analysis by excluding patients with visual acuity worse than 20/25 in the unaffected eye. In this subgroup of 26 patients, responses were significantly lower than in those of the reference group for 8 of the 12 subscales (Table 5). In 2 of the remaining subscales, responses suggested a difference. Responses were not statistically different from patients with diabetic retinopathy in 7 of the 12 subscales. No correlation of subscale responses with visual acuity in the involved eye was present in any of the subscales (Table 6).
The number of systemic medical conditions was examined in all 51 patients to determine if systemic health was related to vision-related QOL (Table 7). There was a strong correlation between the number of systemic conditions and all VFQ-25 subscales except Ocular Pain and Mental Health. Scores for General Health were also compared with the remaining 11 subscales. A significant correlation was seen in General Health responses across all other subscales except Ocular Pain. This remained valid when vision in the better-seeing eye and patient age were controlled in 9 and 10 of the 11 remaining subscales, respectively (Table 7). When the General Health response was controlled, vision in the better-seeing eye again correlated strongly with 9 of the 11 other VFQ-25 subscale responses (Table 8).Last, patient age did not correlate significantly in 9 of the 12 VFQ-25 responses(Table 9), and duration of vision loss did not correlate with any of the responses (data not shown).
This study demonstrates that CVO is associated with a significantly decreased QOL when compared with a reference group without ocular disease. Responses on the VFQ-25 correlated most strongly with visual acuity in the better-seeing eye, the number of systemic medical conditions, and the patients' opinions about their general health. Responses did not correlate with visual acuity in the involved eye.
Patients with CVO reported difficulties with many aspects of daily life. Many patients lived alone and reported that they had difficulty functioning independently. Separate analyses suggest that persons who are not working, representing most of the patients surveyed, have increased difficulty with distance vision, general health, and driving. Scores for patients with bilateral CVO reflect substantial visual morbidity, which can be explained by poor visual acuity in both eyes.
Since VFQ-25 responses correlated with visual acuity in the better-seeing eye, it is possible that the decreased vision-related QOL that we observed was due to the presence of other ocular problems in the uninvolved eye and not related to CVO. We performed subgroup analysis on patients with only unilateral CVO and excellent visual acuity (20/25 or better) in the fellow eye. These 26 patients without measurable visual disability in the fellow eye still demonstrated significantly decreased VFQ-25 scores when compared with the reference group without ocular disease. Since the vision in the fellow eye was excellent, these findings suggest that some additional factor existed to help explain the decreased vision-related QOL.
Patients with CVO often have systemic vascular disease, such as hypertension and diabetes mellitus.11 Since medical comorbidity can affect VFQ-25 scores, 8,9 we correlated number of systemic diseases with VFQ-25 scores and found that this data correlated strongly. A similar strong association was present when patients' responses for the General Health subscale were correlated with other subscales, even when controlling for age and vision in the better-seeing eye (Table 7).
Individuals in our series were significantly older than the comparison groups.3,4 This is likely because CVO typically occurs at an older age than does diabetic retinopathy. The median age in the Central Vein Occlusion Study was 68 years, 12 whereas nearly 50% of patients in the Early Treatment Diabetic Retinopathy Study were younger than 50 years.13 Although VFQ-25 scores may decrease with age, 8 patient age did not correlate well with the VFQ-25 subscale responses in this study.
Vision-targeted QOL surveys have been studied for several ocular diseases. The Visual Function Index (VF-14) has been used to evaluate individuals with cataract, 14 glaucoma, 5 and retinal diseases.15 The VFQ-51 and VFQ-25 have been used to evaluate persons with open-angle glaucoma, cataract, diabetic retinopathy, age-related macular degeneration, cytomegalovirus retinitis, and low vision, 3,4 optic neuritis, 7 and uveitis.9 The VF-14 reports a single value for vision-related function, whereas the VFQ surveys report different subscales. Subgroup analysis may be more useful in evaluating the multiple dimensions of QOL.5 Additionally, the VFQ-51 has a much higher correlation with near and distance vision than the VF-14.3
Generic health-related QOL surveys, such as the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36), have also been used to study ocular diseases. In a study of QOL in patients with glaucoma, the SF-36 was only weakly correlated with visual acuity or visual field impairment.5 The VF-14 and VFQ-51 appeared to be more useful. Other comparisons between the SF-36 and the VFQ-51 show a low degree of correlation, suggesting that they report different QOL dimensions.3
In addition to objective measurement of visual acuity, vision-related QOL questionnaires may reveal an additional dimension of the daily difficulties faced by patients with CVO, as shown in this study. Central retinal vein occlusion is an ocular disease typically associated with a poor visual prognosis.2,16 In our series, the median visual acuity in affected eyes was counting fingers. Although panretinal photocoagulation is recommended for the treatment of anterior segment neovascularization, 17 there is no proven treatment to improve vision in eyes with CVO. Several newer treatments have been reported, including intravitreal triamcinolone acetonide injection, 18,19 radial optic neurotomy, 20 retinal vein cannulation and tissue plasminogen activator injection, 21 vitrectomy surgery, 22,23 injection of intravitreal tissue plasminogen activator, 24- 26 and laser-induced chorioretinal venous anastomosis.27,28 The applicability of surveys in clinical trials has been suggested, 3,4,9,15 and a reliable and easily administered questionnaire, such as the VFQ-25, can report vision-related QOL as an additional outcome measure.
Central retinal vein occlusion remains a challenging ocular disease to treat. This study reminds the physician to remain cognizant of vision in the better-seeing eye, typically not the eye affected with CVO, and the general systemic health of the patient, since these have the greatest effect on QOL in persons with CVO.
Corresponding author: Sharon Fekrat, MD, Duke University Eye Center, Box 3802, Durham, NC 27710 (e-mail: firstname.lastname@example.org).
Submitted for publication November 6, 2001; final revision received March 27, 2003; accepted April 3, 2003.
This research was supported by grant EY11725 from the National Institutes of Health, Bethesda, Md; the Heed Ophthalmic Foundation and AOS–Knapp Fellowship, Cleveland, Ohio (Dr Deramo); the Ronald G. Michels Fellowship Foundation, Ridgewood, Md (Dr Deramo); and Research to Prevent Blindness Inc(Lew Wasserman Merit Award), New York, NY (Dr Lee).
This study was presented in part as a poster at the Association for Research in Vision and Ophthalmology Annual Meeting; May 3, 2000; Ft Lauderdale, Fla.