Objective: To examine the relation between low contrast letter acuity, a new visual function test for multiple sclerosis (MS) trials, and vision targeted health related quality of life (HRQOL).
Methods: Patients in this cross sectional study were part of an ongoing investigation of visual function in MS. Patients were tested binocularly using low contrast letter acuity and Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity (VA) charts. The 25 Item National Eye Institute Visual Functioning Questionnaire (NEI-VFQ-25), 10 Item Neuro-Ophthalmic Supplement to the NEI-VFQ-25, Impact of Visual Impairment Scale and Short Form 36 Health Survey (SF-36) were administered.
Results: Among 167 patients, mean age was 48 (10) years, with median Expanded Disability Status Scale (EDSS) 2.0 (range 1.0–7.5), and median binocular Snellen acuity equivalent (ETDRS charts) 20/16 (range 20/12.5 to 20/100). Reductions in vision specific HRQOL were associated with lower (worse) scores for low contrast letter acuity and VA (p<0.001, linear regression, accounting for age). Two line differences in visual function were associated, on average, with >4 point (6.7–10.9 point) worsening in the NEI-VFQ-25 composite score, reductions that are considered clinically meaningful. Scores for the 10 Item Neuro-Ophthalmic Supplement to the NEI-VFQ-25 also correlated well with visual function. Associations between reduced low contrast acuity and worse vision targeted HRQOL remained significant in models accounting for high contrast VA, EDSS and history of acute optic neuritis.
Conclusions: Low contrast letter acuity scores correlate well with HRQOL in MS. Two line differences in scores for low contrast acuity and VA reflect clinically meaningful differences in vision targeted HRQOL. Low contrast acuity testing provides information on patient reported aspects of vision, supporting use of these measures in MS clinical trials.
Statistics from Altmetric.com
Traditional visual function tests used in multiple sclerosis (MS) clinical trials, such as high contrast visual acuity measured by Snellen charts, have failed to fully detect the often subtle but important visual abnormalities described by patients. Low contrast letter acuity has recently been shown to be a more sensitive measure of both monocular and binocular visual function in MS, even among patients with high contrast visual acuities of 20/20 or better.1–3 Low contrast letter acuity scores correlate well with brain MRI lesion burden,4 visual evoked potentials5 and retinal nerve fibre layer thickness by optical coherence tomography.6 Low contrast letter acuity charts also demonstrated treatment effects in two phase III clinical trials of natalizumab.3 7 8
While these observations support low contrast letter acuity as an outcome measure for MS trials, demonstrating how vision scores may impact functioning and quality of life as perceived by patients is another element that is essential to the evaluation of new outcome measures.9–12 Recently, a Food and Drug Administration draft guidance publication emphasised the importance of patient reported outcomes, including health related quality of life (HRQOL).13 HRQOL scales, such as the 25 Item National Eye Institute Visual Functioning Questionnaire (NEI-VFQ-25) and the Short Form 36 Health Survey (SF-36), are widely used measures of patient reported functioning. These scales are valuable clinically and with regard to their capacity to correlate with new and established clinical endpoints in a variety of disorders, including MS.14–23
The purpose of this investigation was to examine the relation of scores for low contrast letter acuity to vision specific and overall HRQOL in patients with MS. Specifically, we sought to determine whether clinically meaningful differences in binocular visual function (2 line differences) are associated with worse scores for HRQOL, indicating reduced patient reported functioning.
Patients enrolled in an ongoing prospective study of visual outcomes in MS at the University of Pennsylvania participated in this cross sectional study; these data do not overlap with previous HRQOL studies.17 These patients completed HRQOL scales and were not selected specifically based on clinical features or extent of symptoms. The diagnosis of MS was confirmed by standard clinical and neuroimaging criteria.24 Exclusion criteria included a history of acute optic neuritis within 3 months prior to enrolment or a history of ocular comorbidities not related to MS. Patients with binocular Snellen visual acuity equivalents worse than 20/200 were also not included in these analyses given the inability of the low and high contrast acuity charts to provide non-zero scores below this level of acuity without modifying the standardised protocol and testing distance.
Neurological status was determined using the Expanded Disability Status Scale (EDSS) and the MS Functional Composite (MSFC; includes a timed 25 foot walk, nine hole peg test and paced auditory serial addition test).25 Scores for the MSFC in this cohort were standardised based on mean values and standard deviations from a disease free control group of healthy volunteers. Study protocols were approved by the University of Pennsylvania institutional review board, and each subject provided written informed consent. The study was conducted in accordance with HIPAA guidelines.
Health related quality of life assessment
HRQOL instruments were self-administered at the time of the study visit; when possible, patients completed all questionnaires prior to visual function testing. The NEI-VFQ-25 is the standard vision targeted HRQOL scale that is widely used in ophthalmological research and clinical trials using established scoring methods.14 26 The NEI-VFQ-25 consists of 12 subscales; a composite (overall) score is generated as the unweighted average of all items, excluding the single item for general health. Scores for the NEI-VFQ-25 composite and subscales range from 0 to 100, with higher scores indicating better vision specific HRQOL. The NEI-VFQ-25 has been used successfully to demonstrate reductions in HRQOL in patients with MS, optic neuritis and other ophthalmological disorders.15–22 26 While some authors have suggested that 5 point changes in NEI-VFQ-25 scores are to be considered clinically significant,27 28 a more recent study of ophthalmological clinical trial data demonstrated that 4 point differences in composite scores (and 5 point differences in subscale scores) are likely to represent clinically meaningful within person changes for the NEI-VFQ-25.19 20
A 10 Item Neuro-Ophthalmic Supplement to the NEI-VFQ-25 was designed in a previous study by our group to target additional aspects of vision that may not be captured by the NEI-VFQ-25 in patients with MS and other neuro-ophthalmological disorders.17 This new scale distinguished patients with MS from disease free controls to a significant degree in the prior study17 and demonstrated a capacity to capture self-reported visual dysfunction beyond that of the NEI-VFQ-25 alone. Items and composite scores also showed appropriate degrees of internal consistency reliability. The 10 Item Supplement includes 10 items in a Likert scale format similar to that used in the NEI-VFQ-25, each of which is scored on a 0 to 100 scale. A composite (overall) score for the 10 Item Supplement is calculated as the unweighted average of the 10 items. Instructions and content for the 10 Item Supplement have been published15 and were designed based on survey and focus group data for patients with MS, ocular myasthenia gravis and other disorders.
The Impact of Visual Impairment Scale (IVIS) is a 5 item instrument derived from the Functional Capacities Assessment developed by the Michigan Commission for the Blind.29 30 The IVIS captures non-cognitively based difficulties with visual recognition that cannot be corrected with visual aids and is a subscale included in the MS Quality of Life Inventory (MSQLI).29 30 IVIS scores, calculated as the unweighted sum of the item scores, range from 0 to 15, with higher scores representing worse HRQOL. Clinically meaningful changes and differences in IVIS scores have not been established.
The most widely used instrument for measuring general, or overall, HRQOL, the Short Form 36 Health Survey (SF-36), has been used in many MS clinical trials and study cohorts, and represents the core measure in the MSQLI.10 11 23 29–32 The SF-36 includes eight multi-item scales whose scores range from 0 to 100 (higher scores indicate better HRQOL). Summary scale scores for the SF-36, including the Physical Components Summary (PCS) and Mental Components Summary (MCS), are calculated based on standards for the general US population, with a standardised score of 50 (SD 10) representing the mean or “norm.” The SF-36 PCS and MCS have been able to distinguish MS subtype, disease severity and treatment response.10 11 23 29–34 In this study, PCS and MCS scores were used to measure overall aspects of HRQOL; most investigations of the PCS and MCS have established 5 point changes as clinically meaningful.35
Visual function testing
Visual function was assessed using high and low contrast letter acuity testing. High contrast visual acuity was measured using Early Treatment Diabetic Retinopathy Study (ETDRS) charts at 3.2 m (Lighthouse Low-Vision Products, Long Island, New York, USA).36 37 ETDRS charts are the standard used for high contrast visual acuity testing in ophthalmology clinical trials. Low contrast letter acuity was measured using low contrast Sloan letter charts (Sloan charts, 1.25% and 2.5% contrast levels 2 m; Precision Vision, LaSalle, Illinois, USA) (fig 1).1–7 Sloan charts have a standardised format based on that of the ETDRS visual acuity charts (5 letters/line). Each Sloan chart corresponds to a different contrast level, and charts are scored based on the number of letters identified correctly.
All vision testing was performed binocularly (with both eyes open). Binocular testing has been used successfully in recent MS trials and methodological studies,3 4 38 39 and has been recommended as a primary outcome for ophthalmological clinical trials that assess the relationship of visual function and quality of life. Each line on the charts has five letters, and responses by patients are tabulated as the total number of letters read. Clinically meaningful changes in vision scores used herein have been defined in previous studies and trials as 2 line (10 letter) differences. These levels of difference are based on inter-rater and test–retest reliability studies showing that 2 line differences represent changes that are likely to exceed measurement variability.7 8
Calculations and statistical analyses were performed using Stata 10.0 statistical software (StataCorp, College Station, Texas, USA). Linear regression models were used to examine the relation between HRQOL scale summary scores (NEI-VFQ-25 composite, 10 Item Neuro-Ophthalmic Supplement, IVIS overall score and SF-36 PCS and MCS) and visual function scores, accounting simultaneously for age. History of one or more episodes of acute optic neuritis, assessed by self and physician report, was examined as a covariate to explore the effect of optic neuritis history on HRQOL. Additional models incorporating EDSS score and/or high contrast VA score (ETDRS charts) were explored since these factors are likely to independently influence HRQOL scores. Differences associated with p<0.05 were considered statistically significant. To the extent that the present study was designed primarily to provide descriptive exploratory data on the potential relation between visual function and HRQOL scores, using several different preselected scales, we did not adjust for multiple testing. In addition, as recommended by designers of the NEI-VFQ-25 questionnaire, the relation of subscale scores to visual function was explored (using linear regression models) only after significant correlations were noted between vision scores and the NEI-VFQ-25 composite (overall) score, which incorporates 11 of the 12 subscales.
Table 1 presents demographic and clinical data for patients with MS in our cohort (n = 167). Characteristics of gender and age were similar to US MS population statistics.40 With respect to educational background, 123 patients (74%) had completed degrees at college level or beyond. While most patients had relapsing–remitting MS (n = 140, 84%), secondary progressive and primary progressive MS were also represented. Additional psychometric properties of the HRQOL scales for this cohort, including internal consistency reliability, are presented in the supplementary table (available online).
NEI-VFQ-25 composite (overall) scores were reduced among patients with lower (worse) scores for low contrast letter acuity (1.25% and 2.5% contrast levels) and high contrast VA (p<0.001 for all vision tests; linear regression models accounting for age). Detailed analyses of the relation of NEI-VFQ-25 subscale scores and low and high contrast visual acuities are presented in table 2. Two line differences in low contrast letter acuity were associated, on average, with 6.7 point (for 1.25% contrast) and 7.5 point (for 2.5% contrast) differences in NEI-VFQ-25 composite scores (p<0.001). For high contrast VA, 2 line differences corresponded to even greater reductions in NEI-VFQ-25 composite score (10.9 points on average). NEI-VFQ-25 subscales for which 2 line reductions in vision score were most associated with worse HRQOL included those for general vision, near and distance activities, role difficulties and driving (table 2). Accounting simultaneously for high contrast VA scores as well as age in the linear regression models demonstrated that the relation between low contrast acuity and NEI-VFQ-25 composite scores remained significant (1.25% contrast: p = 0.009, 2 lines of vision corresponded to 4.1 HRQOL points, 95% confidence interval (CI) 1.0 to 7.1); 2.5% contrast: p = 0.04, 2 lines of vision corresponded to 4.4 HRQOL points, 95% CI 0.2 to 8.6). Adjusting for EDSS score or history of one or more episodes of acute optic neuritis did not affect the association of low contrast acuity and NEI-VFQ-25 scores (p<0.001 for both 1.25% and 2.5% contrast).
Composite (overall) and item scores for the 10 Item Neuro-Ophthalmic Supplement to the NEI-VFQ-25 were reduced for patients with worse scores for low contrast letter acuity and high contrast VA (table 3); items capturing “unusual eye/lid appearance” were less highly associated with visual function scores in this cohort. Data in table 3 show that 2 line differences in binocular visual function correspond to 4–20 point differences, on average, for the composite and item scores. Supplement item scores that showed the greatest degrees of association with binocular visual function included those for “difficulty parking a car,” “feeling that my two eyes see differently” and “my vision is blurry, not clear or ‘fuzzy.’” Accounting simultaneously for NEI-VFQ-25 composite and high contrast VA in addition to age in linear regression models, the relation between supplement composite scores and low contrast acuity remained significant (1.25% contrast: p = 0.03; 2.5% contrast: p = 0.01).
For the IVIS, higher total scores, in this case indicating worse HRQOL, were associated with lower (worse) scores for low and high contrast acuity (table 2). The median score for the IVIS was 1, compared with 0 for a disease free control group. The correspondence in our study of 2 lines of vision with 1.1–2.0 total score points was significant (p<0.001 for all vision tests, table 2). Adjusting for EDSS score did not affect the association of vision and IVIS scores (p<0.001 for both 1.25% and 2.5% contrast) but when adjusted for high contrast acuity, IVIS scores were not as strongly associated with low contrast acuity (p = 0.02 for 1.25% and p = 0.05 for 2.5% contrast).
In terms of general, or overall, HRQOL, the SF-36 PCS demonstrated lower (worse) scores in the setting of reduced function for both low and high contrast acuity (p<0.001 for all vision tests) (table 2). MCS scores, on the other hand, did not correlate significantly with visual function. Two line differences in low contrast VA scores corresponded to 3 point changes for PCS and 0.3–0.7 point changes for MCS. Accounting simultaneously for EDSS score or high contrast VA reduced the magnitude and significance of the association between PCS and low contrast acuity scores (p = 0.22 for 1.25% and p = 0.34 for 2.5% contrast when accounting for EDSS; p = 0.11 for 1.25% and p = 0.15 for 2.5% contrast when accounting for high contrast VA).
The results of this investigation demonstrate that scores for low contrast letter acuity, a promising visual outcome for MS clinical trials, correlate well with vision specific HRQOL. Patients with clinically significant reductions in low contrast acuity and high contrast VA scores (2 line differences) have worse scores for vision specific HRQOL, with associated differences in NEI-VFQ-25 composite scores representing clinically meaningful decrements in functioning. Data from this study provide important evidence that low contrast acuity scores reflect visual dysfunction in MS beyond that captured by high contrast VA and the EDSS, and indicate that these deficits are associated with clinically meaningful reductions in HRQOL for patients, which is important since patients with MS may have losses of low contrast acuity with relative sparing of high contrast VA.1–7
That high contrast VA had a greater degree of correlation with HRQOL scores is not unexpected as patients are frequently aware of deficits in high contrast VA but may only report subtle symptoms in the context of low contrast acuity loss. While not always as noticeable to patients, reductions in low contrast acuity scores were associated, as shown in our study, with lower (worse) scores for vision specific HRQOL (NEI-VFQ-25 composite), even after accounting for high contrast VA scores. In addition, measures of low contrast vision have been shown to correlate with important “real world” visual tasks, such as driving, reading speed and facial recognition41 42; these tasks may, therefore, be affected in patients with even the most subtle visual symptoms in MS.
The age adjusted regression analyses demonstrate significant associations between low contrast letter acuity scores and physical aspects of HRQOL (SF-36 PCS). MCS scores were not meaningfully related to low contrast VA in this cohort, and that the association of vision with PCS scores was largely accounted for by high contrast VA suggests that the SF-36 may not be as specific an instrument for capturing the impact of subtle visual dysfunction in patients with MS, suggesting that these overall measures of HRQOL are unlikely to be useful in future studies attempting to assess the impact of low contrast visual loss in MS. In previous studies in patients with MS43 or with other neurological disorders, including amyotrophic lateral sclerosis and cervical dystonia,44 45 the PCS and MCS have been shown to lack some desired psychometric features. In our population, there remained limitations of the psychometric properties of these summary scale scores (supplementary table online), further supporting the argument that the PCS and MCS scores should be used and interpreted with caution in this patient population.46
One unique feature of our study is that both the NEI-VFQ-25, the standard HRQOL measure used in ophthalmology clinical trials,14–22 and the IVIS, the vision specific scale included in the MSQLI,29 30 were evaluated simultaneously in an MS cohort. While higher (worse) scores for the IVIS in our cohort were associated with worse scores for low contrast acuity, this association was attenuated when high contrast VA, but not EDSS, was added to the model. IVIS scores thus most likely reflect high contrast VA and may be less sensitive than the NEI-VFQ-25 composite score for capturing more subtle visual dysfunction. Another potential limitation of the IVIS includes a lack of published data regarding the magnitude of difference in the total score that represents a clinically meaningful change. It was thus difficult in our study to evaluate, based on standards, whether or not 2 line differences in vision scores (previously defined for VA as clinically significant)7 8 corresponded to clinically important differences in HRQOL on the basis of IVIS scores alone. Finally, the psychometric profile of the IVIS demonstrates that there is a substantial ceiling effect (see supplementary table online).
In contrast, the NEI-VFQ-25 is an instrument that has been used successfully in many populations (MS and others) to demonstrate reductions in vision specific HRQOL.14–22 27 28 Our data are not only consistent with previous studies that have revealed lower (worse) scores for the NEI-VFQ-25 and its subscales in patients with MS and optic neuritis but are also novel in correlating scores with visual function, particularly low contrast letter acuity, in an MS cohort. Recent reports evaluating the NEI-VFQ-25 have suggested criteria for clinically meaningful changes in scores as ⩾4 points for the composite score and ⩾5 points for each of the 12 subscales.19 20 These changes were reported as criteria for within person changes for NEI-VFQ-25 scores.19 20 In our cohort, 2 line differences in low contrast acuity letter scores were associated, on average, with 7.5 point differences in mean NEI-VFQ-25 composite scores for the 2.5% contrast level, and 6.7 point differences in mean scores for 1.25% contrast, accounting simultaneously for age. Adjusting for EDSS scores as a measure of overall disability, 2 line differences in low contrast acuity corresponded to 6.9 point (2.5% contrast) and 6.0 point differences (1.25%) in mean NEI-VFQ-25 composite scores, suggesting that this HRQOL measure captures vision specific aspects of functioning that are also related to low contrast acuity. In addition to established measures of vision targeted and overall HRQOL, this investigation included exploratory analyses to examine the potential role for a 10 Item Neuro-Ophthalmic Supplement for capturing self-reported visual dysfunction in MS.17 This new scale was developed by our research group based on survey and focus group methods among patients with MS, myasthenia gravis and other neuro-ophthalmological disorders associated with diplopia.15 Further examination of the 10 Item Neuro-Ophthalmic Supplement, particularly in longitudinal studies, is needed to determine its usefulness in MS and other disorders.
At this time, based on data in the present study (see also the supplementary table online) and previous investigations,16 17 the NEI-VFQ-25 appears to be the most promising scale for assessing the impact of visual dysfunction in MS populations. Both the NEI-VFQ-25 and the 10 Item Neuro-Ophthalmic Supplement satisfy important psychometric criteria, including internal consistency reliability (Cronbach’s alpha ⩾0.70, see supplementary table online) and construct validity, as evidenced by significant correlations of scores with visual function tests (tables 2, 3).
The present study did not assess the relation of visual function to HRQOL over time in MS, which also limited our capacity to evaluate some of the psychometric properties of the HRQOL scales longitudinally. To the extent that reduced global HRQOL early in MS predicts subsequent disability at 1 and 5 years following diagnosis, even after accounting for baseline EDSS score,47 48 ongoing investigations will need to examine how scores for low contrast letter acuity, an emerging MS trial outcome, may capture clinically meaningful changes in patient reported visual functioning. Demonstrating a link between visual function scores and established measures of HRQOL, such as the NEI-VFQ-25, is important for ongoing and forthcoming clinical trials of MS and optic neuritis that have incorporated low contrast acuity and ocular imaging measures as exploratory outcomes. This association of visual function and HRQOL will be of particular relevance to trials of neuroprotective agents in which the visual pathways may serve as a model for evaluating efficacy.
Competing interests: None.
Funding: This study was supported by the National Multiple Sclerosis Society Research Grants RG 3208-A-1 and RG 3428-A/2 (LJB), National Eye Institute (NIH) K24 EY 018136 and R01 EY 014993 (LJB). We would like to acknowledge the support of the National Parkinson Foundation Center of Excellence, the McKnight Brain Institute, UF and Shands, and NIH/NINDS K23 NS044997.
Ethics approval: This study was approved by the Institutional Review Board at the University of Pennsylvania.
▸ A supplementary table is published online only at http://jnnp.bmj.com/content/vol80/issue7
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.