In word recognition, eyes rely more on the upper part of a letter than its lower part. The figure shows a sentence and upper and lower halves of words in English (A) and Chinese (B).
eTable 1. Results of the multivariable linear regression model for each score in NEI-VFQ 25 in patients with primary open-angle glaucoma.
eTable 2. Results of the multivariable linear regression model for each score in NEI-VFQ 25 in patients with primary open-angle glaucoma, regarding to superior and inferior hemifields.
Cheng H, Guo C, Chen M, Ko Y, Huang N, Liu CJ. Patient-Reported Vision-Related Quality of Life Differences Between Superior and Inferior Hemifield Visual Field Defects in Primary Open-Angle Glaucoma. JAMA Ophthalmol. 2015;133(3):269-275. doi:10.1001/jamaophthalmol.2014.4908
Previous studies have found that glaucoma is associated with impaired patient-reported vision-related quality of life (pVRQOL) but few, to our knowledge, have assessed how the visual field (VF) defect location impacts the pVRQOL.
To investigate the associations of VF defects in the superior vs inferior hemifields with pVRQOL outcomes in patients with primary open-angle glaucoma.
Design, Setting, and Participants
Prospective cross-sectional study at a tertiary referral center from March 1, 2012, to January 1, 2013, including patients with primary open-angle glaucoma who had a best-corrected visual acuity in the better eye equal to or better than 20/60 and reliable VF tests. The pVRQOL was assessed by a validated Taiwanese version of the 25-item National Eye Institute Visual Function Questionnaire. Reliable VF tests obtained within 3 months of enrollment were transformed to binocular integrated VF (IVF). The IVF was further stratified by VF location (superior vs inferior hemifield).
Main Outcomes and Measures
The association between each domain of the 25-item National Eye Institute Visual Function Questionnaire and superior or inferior hemifield IVF was determined using multivariable linear regression analysis.
The analysis included 186 patients with primary open-angle glaucoma with a mean age of 59.1 years (range, 19-86 years) and IVF mean deviation (MD) of −4.84 dB (range, −27.56 to 2.17 dB). In the multivariable linear regression analysis, the MD of the full-field IVF showed positive associations with near activities (β = 0.05; R2 = 0.20; P < .001), vision-specific role difficulties (β = 0.04; R2 = 0.19; P = .01), vision-specific dependency (β = 0.04; R2 = 0.20; P < .001), driving (β = 0.05; R2 = 0.24; P < .001), peripheral vision (β = 0.03; R2 = 0.18; P = .02), and composite scores (β = 0.04; R2 = 0.27; P = .005). Subsequent analysis showed that the MD of the superior hemifield IVF was associated only with near activities (β = 0.04; R2 = 0.21; P < .001) while the MD of the inferior hemifield IVF was associated with general vision (β = 0.04; R2 = 0.12; P = .01), vision-specific role difficulties (β = 0.04; R2 = 0.20; P = .01), and peripheral vision (β = 0.03; R2 = 0.17; P = .03).
Conclusions and Relevance
Superior hemifield IVF was strongly associated with difficulty with near activities. Inferior hemifield IVF impacted vision-specific role difficulties and general and peripheral vision. The impact of a VF defect on a patient’s pVRQOL may depend not only on its severity, but also on its hemifield location.
Glaucoma is a leading cause of preventable blindness worldwide and is characterized by irreversible optic nerve damage and progressive visual field (VF) loss.1- 3 Although patients with glaucoma usually have good visual acuity until the advanced stage of the disease, many have difficulties in daily life activities. These include reading difficulty, slower reading speed, driving cessation and limitations, bumping into objects, slower walking speed, and a high risk of falls.4- 8
Vision-related quality of life (VRQOL) is a concept that reflects how visual impairment affects one’s personal satisfaction in daily life and one’s functional status.9 Several population-based and clinical studies have evaluated self-reported disability and impaired VRQOL that result from having glaucoma.10- 17 Patients with bilateral glaucoma and greater VF loss have poorer patient-reported VRQOL (pVRQOL) and more difficulties in daily activities including driving, dependence on others, social functioning, mental health, general vision, distance vision, near vision, color vision, and peripheral vision.12,14,15 Only a few reports discuss the impact of the VF defect location on various pVRQOL domains. Sawada et al18 reported that the average mean deviation (MD) of groups of adjacent VF points in the inferior hemifield of the better eye more strongly correlated with pVRQOL than those in the worse eye. Murata et al19 examined whether individual VF test points correlated with pVRQOL and found that important VF test points were clustered along the horizontal meridian. However, these 2 studies did not take nonocular factors into consideration when evaluating the impact of VF defects on pVRQOL. Other studies have shown that nonocular factors, such as race/ethnicity, cognitive ability, education, and comorbid illness, can strongly influence pVRQOL; different conclusions may be reached when a study does not adjust for these covariates.4,20
In this prospective study, we aimed to investigate the impact of VF loss on pVRQOL outcomes in Chinese patients with primary open-angle glaucoma (POAG) while taking potential confounding factors into consideration. Furthermore, we evaluated whether the impact differed depending on the VF location (superior vs inferior hemifield) because mild to moderate glaucomatous VF loss very often involves only 1 hemifield.
This prospective cross-sectional study was conducted from March 1, 2012, to January 1, 2013. Eligible consecutive patients at the Taipei Veterans General Hospital were invited to participate in this study. The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board of Taipei Veterans General Hospital. Written informed consent was obtained from each patient after the aims and methods of the study were fully explained.
All the enrolled patients received a comprehensive ophthalmologic evaluation including a medical history review, assessment of binocular habitual visual acuity (VA) with a Snellen E chart at a distance of 6 m, determination of best-corrected VA, slitlamp biomicroscopy, tonometry, gonioscopy, standard achromatic automated perimetry with the 24-2 Swedish interactive threshold algorithm standard of the Humphrey Field Analyzer 750i (version 4.2, Humphrey Instruments), and dilated ophthalmoscopy. Because we aimed to evaluate pVRQOL in daily life, the presenting VA with the patient’s habitual correction was used for analysis. Measured Snellen VA was then converted to logMAR values for further calculations.
Patients with POAG who fulfilled the following criteria were enrolled: (1) best-corrected VA in the better eye equal to or better than 20/60, (2) intraocular pressure controlled at or below 25 mm Hg, and (3) a reliable VF test within 3 months of enrollment, which was defined as a fixation loss of 30% or lower, a false-positive of 15% or lower, and a false-negative of 15% or lower. Patients with a best-corrected VA in the better eye worse than 20/60 were excluded to avoid confounding effect of low vision on pVRQOL.21 The intraocular pressure limitation was set because we focused on patients with glaucoma whose disease was under control and relatively stable and who had not been experiencing drug switching or laser treatment in the near past. The diagnosis of POAG was made by glaucoma specialists (C.J.-l.L., Y.-C.K., and M.-J.C.) based on reproducible VF defects in a retinal nerve fiber bundle pattern, corresponding glaucomatous optic disc changes, and a normal open angle on gonioscopy examination. A glaucomatous VF defect was defined as a glaucoma hemifield test result outside normal limits or a pattern standard deviation outside 95% of age-specific normal limits. Patients were excluded if they had a patent peripheral iridotomy; a history of laser refractive surgery; a history of laser or incisional glaucoma surgery within 3 months before recruitment; concurrent ocular disease; neurological disease that could cause a VF defect; or severe comorbidities known to affect quality of life, such as dementia, end-stage renal disease, dysfunctional parkinsonism, or systemic malignancy. Patients with prior refractive surgery were excluded because the operation itself may have resulted in decreased sensitivity in the midperipheral VF and had an impact on pVRQOL.22,23
A questionnaire that collected socioeconomic and clinical information was administered at the time of recruitment. The questionnaire asked about the patient’s history of ocular and medical disease, ophthalmic surgery, glaucoma medication class and dosing frequency, educational level, marital status, living (housing) status, employment, and income status. Nonocular diseases were recorded including diabetes mellitus, hypertension, hyperlipidemia, hypotension, migraine, Raynaud phenomenon, coronary artery disease, asthma, chronic obstructive pulmonary disease/chronic bronchitis, liver disease, and renal disease. Psychological disorders were also recorded including depression, anxiety, insomnia, and drug addiction/alcoholism. All psychological disorders had been diagnosed by doctors and medication had been prescribed.
At the time of enrollment, each patient’s pVRQOL was evaluated with the help of a trained assistant using a validated Taiwanese version of the 25-item National Eye Institute Visual Function Questionnaire (NEI VFQ-25 [T]).24,25 This questionnaire aimed to evaluate the self-reported health status of patients with chronic eye disease. It contained 25 vision-related questions categorized to 11 vision-targeted domains and an additional question about the patient’s general health. The score of each domain ranged from 1 to 100, with a higher score indicating better pVRQOL. A composite score was calculated by averaging the scores of 11 vision-targeted domains excluding the general health domain.
An IVF was determined for each participant with POAG using the best sensitivity method.9 From the original achromatic automated perimetry testing with 1 eye at a time, the better raw sensitivity (in decibels) and total deviation value from each pair of spatially corresponding points of 2 eyes of 1 patient were chosen to represent the sensitivity and total deviation of the particular locus in binocular viewing. The 2 most nasal points in each eye were excluded.
The mean of the individual binocular total deviation values was calculated and considered the IVF MD. The IVF was further divided according to location in the superior or inferior hemifield, with respect to the horizontal meridian; the arithmetic mean of the binocular total deviation values of each half was calculated separately.
Descriptive statistics included the mean (SD) for continuous variables and number (percentage) for categorical variables. The scores of each domain of the NEI VFQ-25(T) were skewed and were therefore converted to normal distribution using the SAS procedure PROC RANK, which computed normal scores from the ranks of original values. The association between each demographic, clinical, and socioeconomic variable and each normalized score of the NEI VFQ-25(T) was evaluated initially using univariate analysis. In the univariate analysis, variables with a P value less than .10 were adjusted in the multivariable linear regression model. In addition, habitual VA and the IVF MD were forced to be part of the model where the IVF MD was analyzed as either the full-field IVF MD or as the superior or inferior hemifield IVF MD. Although 13 tests were performed for each variable, the Bonferroni correction method was not adopted to tackle the issue of multiple comparisons because it could also raise type II error. Instead of using P values corrected with the Bonferroni method, we present the original R2, β, and P value of each variable in different models throughout this article. All statistical analyses were performed with commercially available software (SAS version 9.2; SAS Institute Inc).
A total of 193 consecutive patients with POAG were eligible for this study but 3 patients declined to participate and 4 patients were excluded owing to 2 unreliable VF tests after enrollment. Thus, a total of 186 patients with POAG were included in this study. The mean (SD) age was 59.1 (13.1) years (range, 19-86 years), and 131 patients were men (70.4%). Patient demographic and socioeconomic data are listed in Table 1 and patient ophthalmologic characteristics are presented in Table 2.
Multivariable analysis was performed for each domain of the NEI VFQ-25(T) that took relevant demographic, socioeconomic, and ophthalmic factors into consideration based on findings derived from the univariate analysis. The results are shown in eTable 1 and eTable 2 in the Supplement. We found that older age (R2 = 0.27; β = 0.02; P < .001), higher education level (R2 = 0.27; β = 0.16; P = .01), no coronary artery disease (R2 = 0.27; β = −0.74; P = .04), better habitual binocular VA (R2 = 0.27; β = −1.44; P = .04), lower glaucoma medication dosing frequency (R2 = 0.27; β = −0.25; P = .04), and a better MD for the full-field IVF (R2 = 0.27; β = 0.04; P = .005) all correlated with higher composite scores. A β value was defined as the regression coefficient of each variable in the model. Our results showed that the composite score increased by 0.04 units with each 1-dB improvement in MD of IVF (β = 0.04). In addition, 27% variation of the composite scores could be explained by MD for the full-field IVF (R2 = 0.27).
The multivariable linear regression analysis showed that the MD of the full-field IVF was positively correlated with the NEI VFQ-25(T) composite score (R2 = 0.27; β = 0.04; P = .005) and with the scores for near activities (R2 = 0.20; β = 0.05; P < .001), vision-specific role difficulties (R2 = 0.19; β = 0.04; P = .01), vision-specific dependency (R2 = 0.20; β = 0.04; P < .001), driving (R2 = 0.24; β = 0.05; P < .001), and peripheral vision (R2 = 0.18; β = 0.03; P = .02) (Table 3).
After stratifying the IVF into the superior and inferior halves, the MD of the superior hemifield IVF was correlated only with near activities (R2 = 0.21; β = 0.04; P < .001). The MD of the inferior hemifield IVF was positively correlated with general vision (R2 = 0.12; β = 0.04; P = .01), vision-specific role difficulties (R2 = 0.20; β = 0.04; P = .01), and peripheral vision (R2 = 0.17; β = 0.03; P = .03) (Table 4).
In this prospective cross-sectional study of 186 Chinese patients with POAG, we found that the MD of the full-field IVF was associated with nearly half (6 of 13) of the domains of the NEI VFQ-25(T). After stratifying the IVF into the superior and inferior halves, the MD of the superior hemifield IVF was correlated with near activities whereas the MD of the inferior hemifield IVF was associated with general vision, vision-specific role difficulties, and peripheral vision domains.
The full-field IVF was positively correlated with the composite scores and with the scores for near activities, vision-specific role difficulties, vision-specific dependency, driving, and peripheral vision. These findings were in agreement with those of previous studies on patients with glaucoma of different races/ethnicities.12,14,15 However, unlike results from studies of Latino, African American, and white cohorts, the full-field VF was not associated with mental health and social functioning scores in our study.12,14 Several factors influence self-reported mental health and social functioning and their interactions may be complex. For example, Ren and colleagues26 reported that it is culturally unacceptable for Chinese people to use health problems as an excuse to avoid family or social gatherings because of the Confucianism ideology of collectivism. In addition, social activities in Chinese society usually are of the sedentary style. These 2 factors may explain, at least in part, why the social functioning scores in Chinese patients are not significantly correlated with the MD of the full-field IVF.
Early to moderate glaucomatous VF defects usually present with asymmetric distribution between the superior and inferior hemifields. To better understand the difficulties patients with defects that predominantly involve 1 hemifield encounter in daily life activities, we further divided the VF into superior and inferior hemifields to clarify the impact of defect location on pVRQOL. We found that the MD of the superior hemifield IVF was associated with only near activities while the MD of the inferior hemifield IVF had an impact on 3 domains of the NEI VFQ-25(T). In a study with similar aims, Sawada et al18 examined 168 patients with POAG and divided the VF into 10 clusters. They found that the scores of many domains in the NEI VFQ-25 correlated most strongly with the lower paracentral VF cluster in the better eye. Although both studies highlight the importance of the inferior VF in daily life activities, it is difficult to directly compare these 2 studies because of differences in study design and analytic methods. The study by Sawada et al18 did not consider many relevant nonocular factors that can affect pVRQOL. In addition, dividing the VF into many sectors to clarify the impact of VF defect location on pVRQOL may not reflect real-world conditions and the filling-in effect of the brain, in particular.27,28 The filling-in effect helps compensate for gaps in visual perception and is an active visual process that involves creating an actual neural representation of the surroundings rather than merely ignoring the absence of information from the scotoma.27
In the present study, the superior hemifield IVF defect had an impact on the near activities domain of the NEI VFQ-25(T). Similarly, Sawada et al18 found that near activity was most correlated with the upper temporal VF cluster in the better eye. Our findings indicated that patients with binocular superior VF defects had more difficulty reading ordinary print in newspaper, performing work or hobbies requiring close-up vision, and finding something on a crowded shelf. In 1908, Huey29 first proposed that the upper part of a letter or word is more important for perception than is the lower half. The superiority of the upper portion of letters in the Roman alphabet has been demonstrated via psychophysical testing.30,31 Although Chinese characters are quite different from English words, the most important segment of them, which means the ones retaining the character configuration, tends to be located on the top and left side of the word.32,33 Accordingly, the superiority of the upper part in the Roman alphabet may be applied to Chinese characters. The Figure demonstrates that eyes rely more on the upper part of a letter than on its lower part for word recognition.
We found that a worse MD for the inferior hemifield IVF was associated with a lower score for the domains of general vision, peripheral vision, and vision-specific role difficulties. This may explain why patients with glaucoma with worse binocular inferior VF have a slower walking speed, higher rate of falls, and more falls with injury among elderly individuals.6,34 In addition to the MD of the inferior hemifield IVF, we found that having a higher education level (R2 = 0.17; β = 0.12; P = .04) and living with others (R2 = 0.17; β = −0.17; P = .03) were associated with a higher score for peripheral vision (Table 4). This suggests that it may be advantageous to patients with glaucoma to live with families or friends who can provide aid or instructions as needed. Concerning patient safety in performing daily activities, it is important for physicians to remind patients of limitations resulting from VF loss, especially those patients with binocular inferior VF defects.
While we were interested in understanding the impact of VF defects on pVRQOL, we are aware that pVRQOL is affected by many factors other than the VF defect. We found a higher composite score correlated with a higher education level, an absence of coronary artery disease, better binocular VA, lower glaucoma medication dosing frequency, and better full-field IVF. Balkrishnan et al35 found that self-reported difficulty in using eye drops was strongly associated with decreased pVRQOL, as assessed by the NEI VFQ-25. Regarding pVRQOL, physicians may want to simplify the dosing regimen while continuing to preserve the visual function of patients with glaucoma by lowering the intraocular pressure.
This study had several strengths. The glaucoma diagnosis was made by glaucoma specialists and all patients were followed up regularly. Thus, the study results are likely to reflect the real-world situations of patients with glaucoma who have regular follow-up treatment. Stratifying the binocular VF into the superior and inferior hemifields may help physicians understand the difficulties encountered by patients with binocular superior hemifield defects vs inferior hemifield defects and provide more meaningful advice to patients and their families. In addition, we took many other variables that have been reported to influence pVRQOL into consideration while clarifying the impact of VF loss.
Our study findings should also be interpreted in the context of limitations. The sample size was relatively small and all participants were of Chinese race/ethnicity. We included binocular VA and VF as vision-related parameters in the multivariable analysis because VA and VF are 2 factors that are most relevant to pVRQOL; however, we did not include parameters such as contrast sensitivity or stereopsis.21,36 In addition, an IVF was used to simulate binocular VF, which may not fully represent the true VF with both eyes open and has omitted the effect of binocular summation. Nevertheless, several studies have demonstrated that IVF correlates well with other binocular test and assessment of vision.9,37,38 Finally, we used the NEI VFQ-25(T) rather than a glaucoma-specific questionnaire to assess pVRQOL because it is reliable, is often used as a benchmark for comparisons, addresses the issue of pVRQOL comprehensively, and has a validated Taiwan version.24,25,39
The scores of many domains of the NEI VFQ-25(T) were positively associated with the MD of the IVF. Patients with superior hemifield IVF defects had difficulty in near activities that required close-up vision while patients with inferior hemifield IVF defects reported more difficulties with general and peripheral vision and were more likely to report vision-related role difficulties.
Corresponding Author: Catherine Jui-ling Liu, MD, Department of Ophthalmology, Taipei Veterans General Hospital, 201, Section 2, Shih-Pai Road, Taipei 112, Taiwan (email@example.com).
Submitted for Publication: May 14, 2014; final revision received October 12, 2014; accepted October 13, 2014.
Published Online: November 27, 2014. doi:10.1001/jamaophthalmol.2014.4908.
Author Contributions: Drs Liu and Cheng 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: Cheng, Huang, Liu.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Cheng, Guo, Huang.
Critical revision of the manuscript for important intellectual content: Cheng, Chen, Ko, Liu.
Statistical analysis: Cheng, Guo, Huang.
Administrative, technical, or material support: Cheng, Chen, Ko, Liu.
Study supervision: Liu.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.