A Review of Studies of the Association of Vision-Related Quality of Life with Measures of Visual Function and Structure in Patients with Glaucoma in the United States

Abstract

Purpose:

To investigate the association of quality of life (QoL) with ocular structure and function in glaucoma patients, and to identify which aspects of QoL are most closely tied to Visual Field (VF) and Visual Acuity (VA).

Methods:

We conducted a comprehensive review of studies on QoL in glaucoma patients using PubMed, Web of Science, and Google Scholar (from 01–01-1997 to 07–12-2019). A total of 21 studies in the United States that used the 25-item National Eye Institute Visual Function Questionnaire (NEI VFQ) or 51-item NEI VFQ were included. A descriptive analysis of data from the selected studies was conducted. The association between QoL scores and visual function and structure was investigated by ranking the strength of association on a scale from 1 (weakest) to 12 (strongest).

Results:

Studies reported correlations between QoL scores and Visual Structure. Associations were also reported between QoL and Visual Function both cross-sectionally and longitudinally, with stronger association of VF and VA with distance activities (average ranking 9.1 and 9.6), vision-specific dependency (8.7 and 8.9), and driving (8.6 and 9.7). Vision-specific mental health (6.5 and 4.9), vision-specific social functioning (8.4 and 6.2), and vision-specific role difficulties (7.1 and 6.6) domains were more associated with VF than with VA.

Conclusion:

Our study was the first to quantify and rank the strength of association between visual function and QoL domains. Driving and psycho-social QoL domains tended to be most affected by glaucoma-related deterioration of visual function. QoL scores could be used for more patient-centered disease management.

INTRODUCTION

Glaucoma is the leading cause of irreversible blindness worldwide.¹ Primary open-angle glaucoma (POAG) is the most common form of the disease, accounting for the majority (74%) of all glaucoma cases.¹ POAG is defined as an optic neuropathy characterized by optic disc deterioration, retinal nerve fiber layer thinning, and visual field (VF) loss.² In 2015, 57.5 million individuals were estimated to have POAG globally, which is predicted to rise to 65.5 million by 2020.³

Several population-based studies^4–7 and multi-center clinical trials⁸ have reported that glaucoma may lead to a loss of visual structure and function and a decrease in vision-related quality of life (QoL). QoL is increasingly measured in both research and clinical practices as a precise and reliable outcome with strong prognostic value for mortality and poor health outcomes.^9–12 Specifically, glaucoma patients have reported decreases in vision-dependent mobility,¹³ ability to read,^14,15 ability to perform household chores,¹³ and self-perceived visual dysfunction,^4,13–18 even in early stages of disease.^19–22 These patients have also reported increased risk of motor vehicle collisions^13,17 and incidence of falls.¹⁷ Severity, location, and rate of progression of glaucomatous damage are reported to affect the significance of QoL decrease. A more complete understanding of how ocular structure and visual function loss affects QoL is important in guiding patient-centered therapeutic decisions. Moreover, this will aid ophthalmologists in communicating more effectively with patients, possibly improving compliance and outcomes.

Various questionnaires have been used to investigate the effect of glaucoma on QoL. These questionnaires include both generic instruments (i.e. Short Form-36 and the Short Form-12^21,23,24) and vision-specific instruments (i.e. Visual Function Index (VF-14))²⁵ and Visual Activities Questionnaire^8,22). Glaucoma-specific QoL questionnaires used in past studies include the Glaucoma Quality of Life 15- item questionnaire,^19,20 Glaucoma Symptom Scale,²⁶ Symptom Impact Glaucoma Score,²⁷ Glaucoma Health Perceptions Index,²⁷ Glaucoma Symptom Identifier,²⁸ and the Collaborative Initial Glaucoma Treatment Study QoL.^8,27

The vast majority of large glaucoma studies in North America have used the fully validated 25-item National Eye Institute Visual Function Questionnaire (NEI VFQ-25), which is a shorter version of a previous 51-item questionnaire (NEI VFQ-51).^5,29–33 The NEI VFQ-25 measures the influence of visual disability and visual symptoms on generic health domains, including emotional well-being and social functioning, as well as task-oriented domains related to daily visual functioning. The psychometric properties of the NEI VFQ were investigated by Mangione et al³³ in seven tertiary center ophthalmology practices from various geographic regions. This study provided substantial evidence of the validity and reliability of the NEI VFQ when used among persons with various eye conditions, including POAG.^32,33 Although not discussed in this paper, it is important to note that NEI VFQ-25 has also been translated into multiple languages and used in studies such as the Early Manifest Glaucoma Trial (EMGT)³⁴ and the Italian multicenter observational study.³⁵

The present review (i) summarizes the QoL results from the studies in the United States that used NEI VFQ-25 and NEI VFQ-51 specifically for glaucoma patients, (ii) investigates the association of QoL with ocular structure and function in glaucoma patients, and (iii) identifies which aspects of QOL are most closely tied to two specific visual functions – Visual Field (VF) and Visual Acuity (VA).

METHODS

We conducted a comprehensive review of studies in the United States that investigated QoL in glaucoma patients using the NEI VFQ-25 and NEI VFQ-51 questionnaires (Figure 1). Databases including PubMed, Web of Science, and Google Scholar were searched (from 01–01-1997 to 07–12-2019) for relevant articles. Searched keywords included glaucoma, quality of life, questionnaire, 25-item National Eye Institute Visual Function Questionnaire, NEI-VFQ-25, 51-item National Eye Institute Visual Function Questionnaire, and NEI-VFQ-51. Combinations of these keywords with Boolean operators were also used. The reference lists of recently published review articles (2010-present) pertaining to glaucoma and QoL were searched to find additional relevant studies.

Figure 1.

The PRISMA Flow Diagram of Studies’ Identification, Selection, and Inclusion in the Review

All studies evaluating QoL in glaucoma patients were initially included, without regard to whether the instruments were generic, vision-specific, or glaucoma-specific. Only studies conducted in the United States and published as full-text papers in English were considered. Two independent reviewers/UPenn students (AC and MS) reviewed titles, abstracts, and full text articles. Reviewers discussed all selection-related questions with an ophthalmologist/postdoctoral researcher (NK). Among more than 100 records screened, only 21 studies used VFQ-25 and NEI VFQ-51 among glaucoma patients in the United States. All selected studies had sufficiently large sample sizes. We assessed quality of the selected studies using the Newcastle-Ottawa Scale for case-control studies (http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp), a version of the Newcastle-Ottawa Scale modified for cohort studies (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools), and the Cochrane guidelines on Assessing Risk of Bias in a Randomized Trial (https://training.cochrane.org/handbook/current/chapter-08) for the one randomized clinical trial. The Newcastle-Ottawa Scale awards quality points for selection of the study subjects, comparability of groups, and strength of the outcome measure. The Cochrane guidelines address the method of randomization, allocation concealment, masking, use of intention-to-treat analysis, and completeness of follow-up. Only studies considered of good and satisfactory quality (assessed by NK) were included in the review. For more details, see Supplementary Table 1.

The NEI VFQ-25 has the following vision-targeted subscales: (i) global vision rating (General Vision), (ii) difficulty with near vision activities (Near Activities), (iii) difficulty with distance vision activities (Distance Activities), (iv) limitations in social functioning due to vision (Vision-Specific Social Functioning), (v) role limitations due to vision (Vision-Specific Role Difficulties), (vi) dependency on others due to vision (Vision-Specific Dependency), (vii) mental health symptoms due to vision (Vision-Specific Mental Health), (viii) driving difficulties (Driving), (ix) limitations with peripheral vision (Peripheral Vision), (x) limitations with color vision (Color Vision), (xi) ocular pain (Ocular Pain). In addition, it includes a self-rated general health questionnaire (General Health). The composite score ranges from 0 to 100, with 0 representing extreme disability related to vision and 100 representing minimal to no disability.

We conducted a descriptive analysis of data from the selected studies. We compared scores for each QoL domain, and then compared QoL scores between different groups (such as participants with glaucoma and normal controls) within the studies. We also investigated the association between QoL scores and visual function and structure as reported in the selected studies to identify QoL domains with the strongest association with visual function as measured by visual field and visual acuity. Within each study we assigned the rank of the strength of association (absolute value of the correlation coefficient) to each of the subscales, with 1 as the weakest association and 12 as the strongest association. When fewer than 12 subscales were reported, values for the ranks were spaced evenly between 1 and 12. When a study had data from more than one measure of the same visual function such as AGIS score better eye and AGIS score worse eye, we included the measure for the better eye only given the literature saying that better eye has more impact on QoL.^29,36–39 We then averaged the ranks of each domain across all the studies, and weighted the average by the square root of the sample size of the study. Composite scores were excluded from these rankings.

RESULTS

In total, 21 studies were selected for the current review. The study design and results are summarized in Supplementary Table 2.

Overview

Overall, ten studies^{4,5,25,40–46} had cross-sectional design and were based in tertiary university-affiliated ophthalmic practices. Of those, three studies^40,47,48 were multi-central and two studies^4,5 were population-based. Ten studies^{30,32,33,46,48–54} were reports from prospective observational cohorts. One report from a multicenter randomized, controlled clinical trial was also included.⁴⁷

All selected studies included open-angle glaucoma subjects. The vast majority of studies excluded secondary cases of glaucoma and any ocular or systemic diseases that could affect the optic nerve or visual field. All selected studies investigated QoL among adults. The studies by Parrish et al²⁵ and Richman et al⁴² also included patients as young as 15 years old. On average across all studies, 30% of study participants were African American, ranging as low as 8%²⁵ and as high as 66%.⁴¹ Two studies specifically investigated the Latino population, including LALES⁵ and the study in Nogales and Tucson, Arizona.⁴

QoL Scores

Overall, nine studies reported scores for each of the QoL domains among glaucoma patients (Table 1). Of them, four studies^4,5,32,33 compared scores in glaucoma patients versus control groups. The study by Ringsdorf et al⁴¹ compared scores among black versus white populations. The study by Wren et al⁴⁷ compared scores between patients with glaucoma of varying severity, while the study by Lisboa et al⁴⁸ compared POAG patients with normal vision-related quality of life (VR QoL) versus patients with abnormal VR QoL.

Table 1.

National Eye Institute Visual Function Questionnaire-25 Subscales’ Scores in the Studies Conducted in the United States that Investigated Quality of Life in Glaucoma Patients.

Studiesa	Notes	National Eye Institute Visual Function Questionnaire-25 Subscale Scores, Mean (Standard Deviation)
		GH	GV	OP	NA	DA	VSSF	VSMH	VSRD	VSD	D	CV	PV	CS
Gutierrez P, 1997	Reference group		~80.0	~88.5	~87.0	~87.0		~83.5	~86.0		~80.0
	~ Approximately estimated scoresb		~67.5	~81.5	~78.5	~78.5		~71.5	~75.0		~62.5
Parrish RK, 1997		66.4 (20.5)	64.4 (19.0)	77.8 (19.3)	72.8 (25.5)	72.8 (23.3)	84.6 (23.3)	68.2 (24.2)	75.1 (26.4)	82.1 (27.5)	71.3 (24.8)	87.8 (22.0)	69.5 (30.8)	-
Mangione CM, 1998	Reference group	75.0 (17.0)	81.0 (13.0)	90.0 (15.0)	93.0 (10.0)	95.0 (8.0)	99.0 (3.0)	91.0 (11.0)	96.0 (9.0)	99.0 (5.0)	89.0 (14.0)	98.0 (8.0)	97.0 (10.0)	-
		69.0 (19.0)	70.0 (16.0)	89.0 (14.0)	81.0 (22.0)	83.0 (21.0)	91.0 (16.0)	81.0 (18.0)	87.0 (19.0)	93.0 (17.0)	82.0 (18.0)	93.0 (17.0)	76.0 (27.0)	-
Mangione CM, 2001	Reference group	69.0 (24.0)	83.0 (15.0)	90.0 (15.0)	92.0 (13.0)	93.9 (11.0)	99.0 (3.0)	92.0 (12.0)	93.0 (13.0)	99.0 (6.0)	87.0 (18.0)	98.0 (8.0)	97.0 (10.0)	-
		62.0 (25.0)	71.0 (17.0)	89.0 (14.0)	79.0 (23.0)	79.0 (25.0)	79.0 (20.0)	79.0 (20.0)	79.0 (23.0)	79.0 (19.0)	79.0 (28.0)	93.0 (17.0)	79.0 (27.0)	-
Jampel HD, 2002	~ Approximately estimated scoresc	68.1 (26.5)	~71.0	~89.0	~79.0	~79.0	~79.0	~79.0	~79.0	~79.0	~79.0	92.8 (14.7)	~79.0	81.7 (16.2)
Broman AT, 2002	Adjusted difference in scores in glaucoma vs refd	−1.39	−5.14	−3.18	−7.69	−7.88	−7.28	−10.5	−9.58	−13.0	−14.8	−5.19	−8.47	-
Ringsdorf L, 2006	African Americane Whitee	~62.0 ~52.0	~70.0 ~67.5	~87.0 ~82.0	~76.5 ~76.5	~73.5 ~76.5	~88.0 ~87.0	- -	- -	- -	- -	~86.0 ~85.5	~75.0 ~75.5	- -
McKean-Cowdin R, 2008	Reference group	46.3 (1.7)	61.9 (1.3)	73.5 (1.5)	73.2 (1.5)	75.6 (1.2)	87.0 (1.0)	67.0 (1.5)	80.4 (1.5)	80.2 (1.3)	75.4 (1.4)	89.5 (1.1)	81.6 (1.5)	76.8 (0.9)
	POAG patients (VF MD >−2dB)	43.9 (3.9)	62.5 (2.9)	74.1 (3.5)	77.7 (3.4)	79.4 (2.8)	88.2 (2.2)	69.2 (3.5)	88.1 (3.3)	84.8 (3.0)	72.0 (3.2)	91.1 (2.5)	86.7 (3.3)	79.7 (2.1)
	POAG patients (VF MD < −2dB)	47.2 (1.9)	59.3 (1.4)	69 .7 (1.7)	70.8 (1.7)	72.1 (1.4)	86.2 (1.1)	61.9 (1.7)	76.8 (1.6)	73.4 (1.5)	70.6 (1.6)	88.3 (1.2)	77.5 (1.6)	73.2 (1.0)
Wren PA, 2009	POAG patients (VF MD >−2dB)	73.1 (24.2)	86.1(13.2)	93.6 (10.3)	90.6 (13.2)	94.1 (9.4)	98.4 (5.6)	91.8 (11.5)	93.4 (15.7)	97.2 (8.2)	88.6 (18.6)	98.7 (5.5)	92.6 (14.0)	93.3 (6.9)
	POAG patients (−6dB≤VF MD≤−2dB)	67.5 (21.5)	84.6 (12.3)	91.7 (13.4)	88.8 (12.8)	92.1 (11.4)	97.7 (6.7)	91.9 (11.7)	92.5 (15.2)	95.4 (9.5)	90.6 (12.2)	98.1 (8.0)	90.6 (15.3)	92.2 (7.7)
	POAG patients (MD<−6dB)	66.8 (24.7)	78.1 (16.3)	88.4 (15.8)	83.3 (17.7)	86.3 (18.1)	95.3 (11.1)	86.5 (18.2)	89.3 (19.1)	92.8 (17.0)	85.9 (16.6)	96.5 (11.6)	83.3 (19.6)	87.8 (12.3)
Richman J, 2010		-	-	-	-	-	-	-	-	-	-	-	-	75.5 (17.56)
Lisboa R, 2013	POAG patients with normal VRQoL	68.2 (21.4)	81.5 (12.3)	-	89.2 (12.2)	91.8 (10.0)	98.0 (6.2)	91.8 (8.8)	94.6 (11.8)	99.0 (4.4)	89.6 (11.7)	98.3 (6.8)	93.6 (13.0)	92.7 (6.2)
	POAG patients with abnormal VRQoLf	48.4 (24.5)	59.4 (23.5)	-	62.5 (21.4)	72.5 (18.5)	87.9 (16.9)	58.4 (19.0)	57.0 (29.4)	82.0 (25.8)	65.1 (18.9)	89.8 (19.9)	68.0 (27.8)	70.3 (14.2)
Blumberg DM, 2017		-	-	-	-	-	-	-	-	-	-	-	-	89.2g (8.1)

GH=General Health, GV= General Vision, OP=Ocular Pain, NA= Near Activities, DA= Distance Activities, VSSF= Vision-Specific Social Functioning, VSMH= Vision-Specific Mental Health, VSRD= Vision-Specific Role Difficulties, VSD= Vision-Specific Dependency, D=Driving, CV=Color Vision, PV=Peripheral Vision, CS= VFQ-25 Composite Score, N/A=no information is available

POAG=primary open-angle glaucoma, VF=visual field, MD=mean deviation, dB=decibels, VRQoL=vision-related quality of life

^aNine studies^{30,44,46,49–54} from the review that had no information on specific subscales scores were excluded from this table

^bThe scores were estimated from the graph in the manuscript by Gutierrez P, et al, 1997

^cAuthors reported that these scores were similar to those reported in patients with glaucoma in the study by Mangione CM, et al, 2001

^dEffect of glaucoma on NEI-VFQ-25 scores (boot strap estimates), adjusted for presenting VA and demographics. Significant estimates are highlighted in bold print

^eMean scores stratified by race and adjusted for age, number of chronic medical conditions, and Advanced Glaucoma Intervention Study (AGIS) visual field scores. Scores were approximately estimated from a graph in the manuscript

^fAbnormal VRQoL, defined as the presence of a score <50 on any of the subscales, excluding OP

^gMean composite Rasch-calibrated score. All other scores presented in this table were calculated using simple algebraic approach

The majority of studies in this review used a simple algebraic approach for scoring QoL domains. The most recent studies of QoL used Rasch-calibrated⁵⁵ NEI VFQ-25 scores.^{30,43,44,50,52,54} A trend of worse scores was reported for the majority of QoL domains in glaucoma patients versus reference groups.^4,5,32,33 Color vision and ocular pain tended to be the least affected domains when comparing glaucoma patients to reference groups, with a difference in scores ranging from 1 to 5.^{5,32,33,47,48}

QoL and Visual Structure

Several studies reported correlations between QoL and visual structure. A report from DIGS by Gracitelli et al⁵² investigated the correlation between QoL and retinal nerve fiber layer (RNFL) thickness. This study reported that in the multivariable model, each 1-μm-per-year loss in binocular RNFL thickness was associated with a change of 1.3 units per year in the NEI VFQ-25 Rasch calibrated scores. This association remained significant even after adjusting for progressive VF loss (P <0.001).

The study by Daga et al⁵⁴ that included patients with preperimetric glaucoma, perimetric glaucoma (POAG) and healthy controls, reported no significant difference in Rasch calibrated NEI VFQ-25 scores between the preperimetric and control groups (72.8±16.8 vs 73.7±20.2, respectively; p=0.964). However, there was a statistically significant difference in NEI VFQ-25 scores between POAG patients and controls (58.9±18.6 vs 73.7±20.2, respectively; 95% CI of the difference −21.41 to −8.19; p<0.001), as well as between POAG and preperimetric glaucoma groups (58.9±18.6 vs 72.8±16.8, respectively; 95% CI of the difference −21.84 to −5.95; p<0.001).

In the cross-sectional study by Prager et al⁴³ that included 214 eyes of 107 patients with glaucoma, diffuse macular retinal ganglion cell plus inner plexiform layer (RGC+IPL) loss, as measured by SD-OCT, was associated with diminished vision-related QoL. In univariable analyses, patients with diffuse macular RGC+IPL loss had mean composite Rasch calibrated NEI VFQ-25 scores that were 6.15 points lower than the scores of patients with focal damage (ß=−6.15; 95% CI, −11.7 to −0.59; P = 0.03). The effect remained significant even after controlling for mean RGC+IPL thickness (ß=−7.64; 95% CI, −14.2 to −1.03; P = 0.02). However, NEI VFQ-25 scores were not directly associated with RGC+IPL thickness measures.

Another study from the same research group⁴⁶ found that the odds of reporting the “lowest” NEI VFQ-25 score (Rasch-calibrated scores < one standard deviation (SD) below the mean) were 35 times higher in early glaucoma patients with macular damage in the better eye (95% CI, 3.6 to 339.2, P=0.002) and 11 times higher in the worse eye (95% CI, 1.58 to infinity, P< 0.0001) in comparison to early glaucoma patients with no macular damage. Unlike in the study by Prager et al⁴³, there was no detectable difference in NEI VFQ-25 scores between focal and diffuse patterns of loss in worse eyes with macular damage, or in better eyes with macular damage.

QoL and Visual Function

We investigated the association between each of the QoL domains and the main parameters of visual function, VF and visual acuity (VA) (Table 2; Figure 2). Most selected studies reported a correlation between QoL domains and visual function, particularly with VF.

Table 2.

Association between Visual Field (VF) and Visual Acuity (VA) Parameters and National Eye Institute Visual Function Questionnaire-25 Quality of Life (QoL) Subscales

Studya	Parameters	Association Between Visual Field (VF) and Visual Acuity (VA) Parameters and Quality of Life (QoL) Subscales, Correlation coefficient (p-value)													Measure of Association
		GH	GV	OP	NA	DA	VSSF	VSMH	VSRD	VSD	D	CV	PV	CS
Gutierrez P, 1997b	VF, AGIS score,c better eye	−0.16	−0.40	−0.04	−0.31	−0.39	−0.29	−0.28	−0.33	−0.35	−0.46	-	-	-	Spearman rank correlation coefficients
	VF, AGIS score, worst eye	−0.11	−0.36	−0.10	−0.30	−0.35	−0.24	−0.28	−0.32	−0.34	−0.46	-	-	-
	VA, ETDRS score, better eye	0.14	0.35	−0.08	0.27	0.33	0.31	0.18	0.23	0.33	0.33	-	-	-
	VA, ETDRS score, worst eye	0.12	0.35	0.02	0.35	0.32	0.28	0.24	0.25	0.30	0.34	-	-	-
Parrish RK, 1997b	VF,d Binocular visual impairment	−0.06	−0.28	−0.17	−0.28	−0.37	−0.36	−0.32	−0.31	−0.35	−0.22	−0.21	−0.44	-	Pearson correlation coefficients
	VA, AMA score	−0.12	−0.49	−0.09	−0.61	−0.57	−0.49	−0.41	0.49	−0.59	−0.52	−0.47	−0.51	-
Mangione CM, 1998b	VF, AGIS score, better eye	−0.10	−0.21	−0.05	−0.32	−0.29	−0.45	−0.28	−0.27	−0.35	−0.20	−0.12	−0.17	-	Pearson correlation coefficients
	VF, AGIS score, worst eye	0.05	−0.13	−0.02	−0.28	−0.31	−0.37	−0.24	−0.31	−0.33	−0.24	−0.21	−0.19	-
Mangione CM, 2001	VF, AGIS score, better eye	−0.13	−0.29	−0.06	−0.40	−0.42	−0.47	−0.36	−0.23	−0.51	−0.24	−0.15	−0.25	−0.41	Pearson correlation coefficients
	VF, AGIS score, worst eye	0.00	−0.22	−0.05	−0.36	−0.40	−0.33	−0.29	−0.25	−0.35	−0.26	−0.23	−0.23	−0.37
Jampel HD, 2002	VF, Esterman score	0.23	0.30	0.25	0.26	0.39	0.42	0.26	0.37	0.32	0.24	0.35	0.45	0.44	Partial correlation coefficientse
	VF, MD	0.31	0.37	0.25	0.29	0.45	0.41	0.37	0.50	0.24	0.35	0.31	0.47	0.49
Ringsdorf L, 2006f	• African Americans														Spearman partial (adjusted for age, sex, and VA) correlations
	VF, AGIS score, better eye	−0.08	−0.19	−0.07	−0.23	−0.20	−0.25	-	-	-	-	−0.19	−0.20	-
	VF, AGIS score, worst eye	−0.04	−0.21	−0.05	−0.17	−0.14	−0.11	-	-	-	-	−0.14	−0.29	-
	• Whites
	VF, AGIS score, better eye	−0.05	−0.26	−0.09	−0.27	−0.35	−0.29	-	-	-	-	−0.18	−0.30	-
	VF, AGIS score, worst eye	−0.05	−0.23	−0.03	−0.20	−0.22	−0.20	-	-	-	-	−0.17	−0.20	-
McKean-Cowdin R, 2008	Visual field loss (VFL, MD≤−2 dB), better eye	0.32	0.35	0.47	0.59	0.46	0.07	0.70	0.73	1.14	1.50	0.65	−0.01	0.53	Linear regression β-coefficientsg
	VFL, worst eye	0.17	−0.05	0.12	0.23	0.21	−0.11	0.53	0.34	0.71	0.86	0.16	0.34	0.27
Wren PA, 2009	VF, MD, better eye	0.16	0.35	0.18	0.23	0.31	0.26	0.26	0.24	0.30	0.29	0.23	0.22	0.37	Pearson correlation coefficients
	VF, MD, worst eye	0.12	0.32	0.13	0.18	0.25	0.20	0.20	0.19	0.19	0.23	0.21	0.26	0.30
	VA, ETDRS score, better eye	0.12	0.32	0.12	0.17	0.21	0.12	0.15	0.23	0.16	0.24	NS	0.11	0.26
	VA, ETDRS score, worst eye	0.18	0.36	0.11	0.29	0.36	0.14	0.21	0.31	0.28	0.30	0.30	0.21	0.38
Richman J, 2010	VF, Esterman score	−0.08	−0.42	0.19	0.50	0.43	0.45	0.37	0.48	0.45	0.31	−0.32	−0.42	0.53	Spearman rank correlation coefficients
	VF, Integrated score	0.03	0.40	−0.26	−0.47	−0.39	−0.45	−0.35	−0.49	−0.46	−0.36	−0.30	−0.43	−0.53
	VA, Binocular	0.15	0.47	−0.24	−0.46	−0.44	−0.41	−0.36	−0.45	−0.44	−0.33	0.16	0.30	−0.48
Ekici F, 2015	VF, MD, better eye	0.11	0.23	−0.01	0.20	0.30	0.29	0.17	0.14	0.24	0.31	0.26	0.31	0.29	Spearman Correlation Coefficients
	VA, LogMar, better eye	−0.14	−0.16	0.03	−0.15	−0.23	−0.18	−0.13	−0.12	−0.21	−0.24	−0.14	−0.10	−0.18
	VF, MD, worse eye	0.11	0.28	0.12	0.25	0.31	0.35	0.24	0.30	0.25	0.24	0.26	0.38	0.35
	VA, LogMar, worse eye	−0.08	−0.36	−0.03	−0.28	−0.30	−0.29	−0.23	−0.32	−0.26	−0.24	−0.16	−0.30	−0.31
Sun Y, 2016	VF, MD, better eye	0.11	−0.01	−0.01	0.20	0.30	0.29	0.16	0.14	0.24	0.31	0.26	0.31	0.29	Spearman Correlation Coefficients

GH=General Health, GV= General Vision, OP=Ocular Pain, NA= Near Activities, DA= Distance Activities, VSSF= Vision-Specific Social Functioning, VSMH= Vision-Specific Mental Health, VSRD= Vision-Specific Role Difficulties, VSD= Vision-Specific Dependency, D=Driving, CV=Color Vision, PV=Peripheral Vision, CS= VFQ-25 Composite Score, AGIS= Advanced Glaucoma Intervention Study; ETDRS=Early Treatment Diabetic Retinopathy Study; AMA= American Medical Association

POAG=primary open-angle glaucoma, VF=visual field, MD=mean deviation, dB=decibels, QoL=quality of life

^aTen studies^{4,30,43,44,46,48,50–52,54} from the review that had no information on correlation coefficients to measure association between VF and VA with QoL were excluded from this table

^bVFQ-51 questionnaire was used. Table includes only domains comparable with VFQ-25.

^cHigher AGIS scores represent greater field loss

^dVF scores were adjusted for VA

^ePartial correlation coefficients were calculated based on multiple linear-regression models

^fVFQ-25, excluding VSMH, VSRD, VSD, and D domains due to time constraints

^gLinear regression β-coefficients were used to measure the association between VFL and adjusted mean NEI-VFQ-25 scores. QOL scores were adjusted on age, gender, education, employment status, income, acculturation, health insurance, vision insurance, number of comorbidities, knowledge of glaucoma status, and VA

Figure 2.

Association between Visual Function and Quality of Life in Glaucoma Patients from 11 reports.

Visual function was measured by Visual Acuity and/or Visual Field. Within each study, we assigned the rank of the strength of association (absolute value of the correlation coefficient) to each of the subscales, with 12 as the strongest association and 1 as the weakest association. When fewer than 12 subscales were reported, values for the ranks were spaced evenly between 1 and 12. When a study had data the better eye and the worse eye, we included the measure for the better eye. We then averaged the ranks of each domain across all the studies, and weighted the average by the square root of the sample size of the study. Composite scores were excluded from the rankings.

For example, LALES reported a correlation between severity of VF loss and most NEI-VFQ-25 subscale scores.⁵ This correlation persisted even after controlling for knowledge regarding POAG diagnosis.⁵ Researchers at Columbia University also found that composite NEI VFQ-25 score was associated with both binocular 24–2 (β = 1.95; 95% CI, 0.47–3.43; P = 0.01) and 10–2 (β = 2.57; 95% CI, 1.12–4.01; P=0.001) sensitivities, with β representing the VF (dB) difference associated with a unit change in the Rasch-calibrated scores.⁴³ However, the 10–2 VF in a univariable model showed an almost two-fold better fit to the data (R² = 9.2% vs 4.9%). In the multivariable model, when both 10–2 and 24–2 VFs were included, the association of QoL with binocular 24–2 VF was no longer observed (P = 0.89), but QoL remained associated with the binocular 10–2 VF (β = 2.70; 95% CI, 0.36–5.04; P = 0.02).

All studies reported that both VA and VF had the weakest association with general health and ocular pain domains of QoL. For VF, the average rank across studies of association with general health was 2.1, and with ocular pain was 2.1. For VA, the average rank of association with general health was 3.6 and with ocular pain was 2.6. The color vision domain also tended to be less associated with both VF and VA (average rank 5.2 and 4.3, respectively). The peripheral vision domain was among the least associated with VA (average ranking 2.1), but did show a stronger association with VF (average ranking 7.1). The majority of studies reported stronger association of both VF and VA with distance activities (average ranking 9.1 and 9.6), vision-specific dependency (average ranking 8.7 and 8.9), and driving (average ranking 8.6 and 9.7). Vision-specific mental health (average ranking 6.5 and 4.9), vision-specific social functioning (average ranking 8.4 and 6.2), and vision-specific role difficulties (average ranking 7.1 and 6.6) domains were more associated with VF than with VA. In contrast, general vision (average ranking 6.3 and 8.4), near activities (average ranking 6.9 and 8.7) were less associated with VF than with VA.

QoL and Longitudinal Changes in Visual Function

Several studies reported significant correlations between rates of progressive VF loss and longitudinal changes in Rasch-calibrated QoL scores. During the follow-up period in the Diagnostic Innovations Glaucoma Study (DIGS),³² each 1-dB change in binocular standard automated perimetry (SAP) mean sensitivity (MS) per year was associated with a change of 2.9 units per year in the NEI VFQ-25 Rasch-calibrated scores (R²=26%; P < 0.001). Eyes with more severe disease at baseline were also more likely to have lower QoL scores during follow-up. Similar results were reported by Abe et al using frequency doubling technology (FDT) perimetry.⁵¹ However, the multivariable model containing baseline and rate of change information from SAP had a stronger ability to predict change in NEI VFQ-25 scores compared to the equivalent model for FDT (R² of 50% and 30%, respectively; P = 0.001). In another report from DIGS, Abe et al reported a significant correlation between the change in the NEI VFQ-25 Rasch scores during follow-up and change in different regions of the VF.⁵⁰ The association for central inferior VF area was significantly stronger than those of central superior area (P = 0.01) and peripheral superior area (P = 0.001), but not peripheral inferior area (P = 0.17).

DISCUSSION

The present review summarizes the results from 21 studies of POAG patients conducted in the United States. All studies investigated the association of visual function and/or structure with QoL as measured by the NEI VFQ-51 or NEI VFQ-25 questionnaires. Our goal was to gain a better understanding of how this progressive disease affects the quality of life of patients, which will aid ophthalmologists in understanding how to best communicate with patients and manage the disease.

As expected, all selected studies reported a negative impact of glaucoma on QoL scores. This association persists even after adjustment for VA.⁵ Moreover, research findings suggest that adults with glaucoma experience a measurable loss in QoL scores early in the disease with little loss in visual function; this decline continues with increasing disease severity.⁶

Our study was the first to quantify and rank the strength of association between visual function and QoL domains. This approach provides more specific information on how glaucoma affects QoL, allowing ophthalmologists to better anticipate and meet patient needs. POAG appears to impact the majority of QoL domains, even at early stages of the disease. The domains least associated with visual function were ocular pain and color vision, which is explicable since they are not typical symptoms of POAG. Interestingly, general health was also among the least associated domains. Glaucoma has previously been associated with a number of systemic diseases, including hypertension, diabetes, obesity. This lack of association with the general health domain requires further research.

Our review revealed that driving was one of the most affected domains of QoL. This is in line with other studies that reported the greatest impact of glaucoma-related loss of visual function on driving difficulty. For example, Freeman et al⁵⁶ found that various measures of visual function were associated with driving modification after two-year follow-up. In a multivariable model, worse baseline VA was related to reduction in mileage, and lower peripheral fields were related to cessation of night driving. The association between visual function loss and driving could be explained by limitations in glare and dark adaptation, as light scatter can aggravate loss of visual function in glaucoma patients.^19,56 One study reported that difficulties with bright lights and with light and dark adaptation were the most frequently reported visual symptoms, while visual distortion was the most bothersome.²⁷ Another study also showed that the majority of patients reported problems with glare and adaptation to different levels of lighting.²⁰

In addition to driving, our review found that vision-specific mental health, vision-specific role difficulties, and vision-specific dependency were among the most affected domains of QoL. The impact of glaucoma on psychosocial function has historically not been clear. LALES reported that vision-related QoL is greatly impacted by severe VF loss from glaucoma, even among individuals previously unaware of their diagnosis.⁶ Studies of other eye diseases (including cataract, glaucoma, and macular degeneration) show that the presence of these diseases does not generally affect QoL after adjusting for other important variables.⁵⁷ These results suggest that merely having a glaucoma diagnosis may not be as impactful to QoL as having physical problems and symptoms, such as blurred vision and decreased vision.

The majority of studies reported an association between visual function and QoL, both cross-sectionally and longitudinally. VF loss was more strongly associated with social and psychological dimensions of QoL, while VA was more associated with functional dimensions of QoL. Severity of VF loss at baseline and faster rate of change were both associated with larger declines in QoL scores.³⁰ It is possible that patients with slow-progressing glaucoma have sufficient time to develop compensatory strategies to cope with failing functional status. In contrast, fast progressors may have less time to adapt to substantial declines in daily activities, contributing to poorer QoL. In particular, loss of VF in the inferior region (versus the superior region) had a strong association with lower QoL scores.^50,58 Loss of vision in the inferior region has a greater impact on the ability to perform daily activities, such as reading, walking down stairs, or viewing peripheral objects while walking.⁵⁸ It also has been associated with a higher rate of falls and more falls resulting in injuries among elderly individuals.^59,60

Glaucoma-related structural damage, defined as RNFL thickness loss⁵⁴ and RGC+IPL loss⁴⁴ was also strongly associated with decline in QoL in POAG patients. Moreover, there are reports that significant neural losses may occur in the macular area of patients with glaucoma, which are likely to be relevant to several domains of QoL.⁶¹ However, unlike measurements of RNFL thickness⁶² and patterns of RGC+IPL loss,⁴⁴ these changes are typically not detectable by the standard 24–2 pattern of VF testing.⁴³ These studies demonstrate that assessment of structural damage could provide additional information for predicting change in QoL besides what can be gathered by assessment with standard perimetry. Patients with structural damage should be closely monitored and may require more intensive treatment to prevent or delay the decrease of QoL.

Our study is the most up-to-date review of the research that investigates QoL in glaucoma patients in the United States from 1997 to present. We excluded reports from other countries in order to include a patient population with the same standards and guidelines of glaucoma diagnostics and management. While there is a close similarity in guidelines in glaucoma management in other industrialized countries, there are differences in health systems, which could impact the management of the disease, such as health insurance coverage, cost of treatment, accessibility, and availability of services. While NEI VFQ-25 has been translated into multiple languages, there are inter-country cultural nuances and specifics that could possibly affect patients’ responses to the QoL questionnaire. In our review, we tried to minimize the impact of these factors and to include the most “combinable” studies in our analysis. All selected studies used scientifically sound methodology and included sufficiently large sample sizes. The most recent studies included racially diverse populations, providing information of QoL among African American and Latino populations.

While we considered including studies that used other validated questionnaires, we decide to use only NEI VFQ data in order to compare and quantitatively combine the results from these studies in the analysis. Since the vast majority of large glaucoma studies in the United States used the NEI VFQ-25, our review was able to include a relatively large sample of studies, which would not be possible if we decided to work with data from other vision-specific questionnaires. Moreover, the NEI-VFQ is reported to be sensitive to changes in visual acuity and visual field. This was particularly important since our review explored the association of vision-related QoL with measures of visual function and structure.

To our best knowledge, we were the first study to rank the strength of association between each QoL domain and visual function. Moreover, we included the most recent studies that investigated association of QoL with parameters of visual structure as measured by SD-OCT, thus providing a comprehensive review of our topic.

There are several limitations to this review. First, we included studies that used NEI VFQ-25 and NEI VFQ-51 questionnaires, which are not glaucoma-specific questionnaires. Consequently, they do not cover certain dimensions of visual impairment that can be relevant in glaucoma patients, including contrast sensitivity, stereo acuity, glare sensitivity, and dark adaptation. However, unlike some glaucoma-specific instruments, such as the Glaucoma Quality of Life-15 (GQL-15) questionnaire,⁶³ the psychometric properties of NEI-VFQ-25 have been evaluated. It is a reliable and valid instrument and is one of the most commonly used instruments in glaucoma research. Second, patients in the majority of the selected studies were mostly aware of their glaucoma diagnoses while responding to the QoL questionnaires. Moreover, it is possible that some of these patients were aware of the severity of their disease. This knowledge may have influenced how they responded to the questions, thus impacting their QoL scores. However, our review did include a report from LALES that investigated patients (75% of the whole sample) before they were diagnosed, minimizing the response bias. Third, this review only included studies conducted in the United States, which may limit its generalizability to other populations. Finally, the majority of the studies in this review used a simple algebraic approach for scoring QoL domains. This traditional summary scoring hypothesizes that all questions have equal importance, without statistical justification of rating scales and scoring systems.⁶⁴ However, the most recent studies of QoL included in our review used Rasch-calibrated⁵⁵ NEI VFQ-25 scores,^{30,43,44,50,52,54} thus addressing the unidimensionality issue within a scale or subscale. The Rasch model use the responses of the subjects to define the item difficulty of each question. Calibrating responses in different categories allows for a more valid scale, compared with a ‘one size fits all’ scoring system.³⁰

In conclusion, this review showed a trend of worse scores for the majority of QoL domains in glaucoma patients. Driving and psycho-social QoL domains tended to be most closely tied with glaucoma-related deterioration of visual function (VA and VF respectively) Since QoL scores were associated with objective functional and structural measurements of glaucoma, they could be used for more patient-centered management of disease. Further research is needed to develop a more glaucoma-specific quality of life measurement, and to include it in a routine evaluation of glaucoma patients.