The elderly population is increasing in the U.S. and other Western countries at a dramatic rate. According to the U.S. Census 2000,1 35 million Americans are aged ≥65 years. This represents 12.4% of the population compared with only 4.1% aged >65 years 100 years ago. The elderly population is aging and expanding in numbers. Those aged >85 years are growing at three times the rate of the general population. As many as 50,000 Americans are now aged >100 years. More than 95% of those aged >65 years live in the community and not in nursing homes.1 Maintaining vision is an important factor in maintaining the ability to live independently.
Until about 10 years ago, only high contrast acuity was well documented in population studies involving this age segment.2–4 Few studies included other measures of vision function, such as contrast sensitivity, disability glare, vision in reduced illumination, or color vision. More recent population studies have included some of these other measures, primarily contrast sensitivity and disability glare.5 Many previous studies using small samples have shown significant losses of contrast sensitivity with increasing age6 and increases in disability glare with age.7
Conditions of reduced light level or reduced contrast or the presence of glare are expected to impact vision function more severely in elderly persons given the well-known changes in the eye with normal aging and previous publications on normal age changes in vision function. Pupil size decreases in old age, particularly in dim light;8 light transmission through the eye decreases in a spectrally selective way such that blue light is poorly transmitted;9 and scatter increases primarily in the lens but also in the retina.10, 11 The result is reduced retinal image contrast and reduced retinal illumination and increased susceptibility to glare. The changes in the ocular media can also occur asymmetrically, potentially affecting binocular vision and stereopsis. In addition, adaptation in the retina is slowed.12 Thus, vision function is expected to be worse in low contrast and low light conditions and NOT in the high luminance and high contrast conditions that are normally used to assess vision clinically. It also is expected that vision function in the presence of continuous glare and after exposure to a brief glare source is significantly affected in the older eye even in the absence of clinical disease.
The purpose of the Smith-Kettlewell Eye Research Institute (SKI) Study was to assess vision function comprehensively in an older population using a battery of tests emphasizing low contrast function. In addition, we wanted to evaluate the relationship between vision function and visual performance, as well as the relation between vision and subjective complaints and vision and general health and functioning. Because we wanted to relate to visual complaints in daily life, all vision tests were done binocularly with habitual correction. These goals were achieved using a cross-sectional design.
The longitudinal phase of the study was designed to determine whether nonstandard vision tests, specifically low contrast tests, could predict future loss of visual acuity. The SKI study was conducted at the Buck Center for Research in Aging in Marin County, California in collaboration with John A. Brabyn, Marilyn E. Schneck, Lori A. Lott, and Arthur Jampolsky from the Smith-Kettlewell Eye Research Institute.
The SKI Study Population
The SKI study population consists of a random sample of community-living elderly persons from Marin County, California. Initially 900 observers were tested. The average age at the first visit was 75.5 years (range, 58 to 102 years; SD, 9.3 years). The lower limit for eligible age at the time of recruitment was 55 years. The population was deliberately oversampled in the oldest age groups. Three hundred people were aged >80 years at the first visit. Details about the population can be found in Haegerstrom-Portnoy et al.13 The population was invited for longitudinal follow-up evaluation, and 596 observers were retested on average 4.4 years later (SD, 1.0 years), and 451 were retested a third time on average another 2.6 years later (SD, 0.24 years). This is primarily a white sample with high socioeconomic status and high educational level with >50% of participants having finished college. Men and women were nearly equally represented initially (women, 53%), and women represented 57% of the survivors at the third test. There were no exclusion criteria, which means that some of the participants had ocular disease.
Vision Function Measures
Table 1 lists all the vision functions tested. Methodological details can be found in Haegerstrom-Portnoy et al.13 All the vision functions other than Amsler grid were tested binocularly with habitual correction. In addition, a vision questionnaire and a general health and functioning questionnaire were used at all three test times. Physical measures were also included.14 Information about driving was gathered as well.15 Face recognition and performance on a search task for grocery items were added at the second and third test times.
SKI Study: Cross-sectional Vision Results
To provide an overview of the age changes in this large battery of vision functions, a metric is needed that allows comparisons between measures with different units. Acuity measures are in logarithm of the minimum angle of resolution (logMAR); glare recovery is expressed in seconds; stereopsis is measured in seconds of arc; critical flicker/fusion frequency (CFF) is expressed in hertz; field area is measured in square degrees, and so on. We chose to express the results of the older participants’ performance in terms of the number of times worse compared with young observers (our own results on the same tests given to young observers; see Fig. 1 caption for values). Fig. 1 shows the times worse for all the vision functions for the different older age groups (in 5-year bins). The results reflect the medians. The stereo results are categorical and are plotted at the intermediate values of the plates. It is clear from Fig. 1 that each vision measure has its own “aging function.” Some measures show little change during the 4 decades, whereas others demonstrate tremendous losses. Each vision function starts “falling off” at different ages. The loss for the oldest age group (>90 years) ranges from 1.2 times worse for CFF to nearly 18 times worse for low contrast acuity in glare (disability glare) corresponding to a visual acuity of 20/720. Standard high contrast visual acuity, which is the measure most often used in clinical practice, is insensitive to age changes in this randomly selected population, showing only 2.4 times worse values for the oldest age group (>90 years). The three functions showing the worst losses are disability glare, glare recovery, and attentional visual field area. Detailed results in the original units for each vision measure have been previously published.13 Even though each vision function appears to have its own “aging function,” we have previously shown that the same-shaped exponential function with a time constant of ∼15 years fits all the spatial vision measures (all the acuities and contrast sensitivity), provided that the functions can be moved along the x axis to take into account the different ages when they begin to fall off and the difference in baseline values for young observers (sliding on the vertical axis). For example, low contrast acuity in glare begins to fall off about 12 years before standard high contrast visual acuity.13
Spatial Vision Measures
To develop a more intuitive “feel” for the relative changes in spatial vision function, Fig. 2 presents several letter m’s of arbitrary sizes but sized in proportion to the median acuity for four spatial vision measures for two age groups, the 60 to 65 age group and the 85 to 90 age group. Similar results limited to those with 20/40 or better high contrast distance visual acuity have been previously published.16 All the observers in the two age groups are included here. The values for the oldest age group would not fit on the page. The median size for high contrast targets in bright light differ by a factor of 1.8 for the two age groups, whereas the median size for low contrast letters in glare is 3.7 times worse for the 85- to 90-year age group than for the 60- to 65-year age group. Within an age group, the different measures increase in threshold size 3.3 times for the younger group and 6.75 times for the older group as the testing situation is changed. The increased variability with age is not demonstrated by either Fig. 1 or Fig. 2 but is important. We have previously shown that for a person, performance on nonstandard vision tests cannot be predicted from measures of visual acuity.17 Thus, even though the low contrast measures are correlated with high contrast visual acuity, the measurements are not redundant. Two observers can have identical high contrast visual acuity and widely different acuity in the presence of glare, for example.
Vision Function for Those with 20/40 or Better High Contrast Acuity
Even if only those persons are included who would be able to pass a Department of Motor Vehicles (DMV) acuity screening (high contrast acuity of 20/40 or better), a large proportion have vision function 10 times worse than a young “normal” person on nonstandard tests. Fig. 3 shows the percentage of those with 20/40 or better high contrast visual acuity who are 10 times worse than “normal” on four measures of vision function. In the oldest age group in this figure (those >85 years), about 55% have vision function reduced >10 times from a young “normal” person when trying to read low contrast letters in the presence of disability glare (>20/400; solid diamonds), and about 40% have gross stereopsis worse than 10 times that of a young “normal” person (340 sec arc; open circles). About 40% have glare recovery times >90 s (open squares) even though they were only exposed to the bright light source (3300 cd/m2) on the Berkeley Glare test for 1 minute. In addition, nearly 25% have acuity reduced >10 times from that of a young “normal” person when reading low contrast letters in reduced illumination (>20/400; solid squares). Thus, even though all these observers would pass a DMV vision screening and are classically not considered visually impaired, many have severely compromised vision function in nonideal conditions. Reduced target contrast, presence of glare, and poor illumination are conditions commonly encountered in daily life, including while driving. Our results on driving self-restriction indicate that these vision function changes contribute significantly to changes in driving behavior in elderly persons.15 For example, people with poor low contrast acuity in the presence of glare more often restricted their driving to daytime hours even though their high contrast acuity was the same as those who did not restrict their driving.
Simulation of Aging Losses
These large changes in nonstandard vision functions have implications for functioning in daily life. To grossly simulate the impact of loss of contrast sensitivity alone, a plastic sandwich bag was placed in front of the camera. This reduced contrast sensitivity by ∼0.4 log units for young observers on the Pelli-Robson chart, which mimics the median contrast sensitivity of the 80-year-old persons on this test. The simulation minimally affected high contrast acuity. Fig. 4. A and B show examples of a street scene with and without the contrast reduction and a face against a bright window (Fig. 4. C and D). In the street scene, details like the buckling of the sidewalk and the area of dirt and leaves on the right become obscured, whereas the face becomes unrecognizable with the simulated contrast loss despite the small and peripherally located window that acts as a glare source. Other photographic examples can be found in Brabyn et al.18 This gross simulation only addresses one aspect of the age-related losses. Of course, in real life for the median 80-year-old person, not only is contrast sensitivity reduced but also adaptation is dramatically delayed, retinal illuminance is significantly decreased, the color appearance of objects is changed causing color discrimination losses, stereopsis is decreased, attentional visual field is significantly decreased, and visual acuity is decreased by a small amount. For a given person, the changes in each of these functions will differ from the population results, and the person’s performance on these various measures cannot be predicted from his or her performance on the standard acuity chart. Many elderly persons are not themselves aware of these large losses in vision function, relying instead on their doctor’s statement that their vision is “fine” for their age (based exclusively on high contrast high luminance visual acuity).
Other Vision Function Changes
Vision functions other than spatial vision measures also show dramatic losses with age. We have previously shown that gross stereopsis declines surprisingly and significantly with age.19 Only 55% of the population of 900 elderly persons were able to pass the 85 sec arc disparity plate on the Frisby stereotest, and only 41% of those >75 years were able to pass this fairly gross level of stereopsis. As few as 25% of those >85 years demonstrated 85 sec arc of gross stereopsis. The loss of stereopsis, we argued, is mostly caused by ocular changes without having to invoke cortical losses.19 By eliminating those observers with poor ocular functions (prolonged glare recovery, poor disability glare scores, and abnormal Amsler grid scores), the pronounced age effect found for stereopsis disappeared, suggesting that abnormalities within the eyeball were responsible for the poor results on the Frisby stereotest.
Another function that collapses with age is the attentional visual field. The attentional visual field was tested in a Synemed light-emitting diode perimeter (Benicia, CA; see Haegerstrom-Portnoy et al.13 for details). Initially, for the “standard” field test, the participant was asked to fixate a red fixation target and push a button when the green peripheral targets were flashed randomly at eight locations along five meridia (60°, 185°, 225°, 315°, and 355°). For the attentional field test, the task was to count the number of times the fixation target light went “on/off” while pushing the button when the peripheral targets were presented. The central and peripheral targets were of high intensity and were suprathreshold. The most peripheral target was at 70° eccentric for four of the meridia and 55° for one meridian. The maximum average radius was 67°. The standard task was corrected for spurious responses, whereas the attentional field was corrected for spurious responses and the percent correct for the central attentional task. Total test time for the attentional component was ∼3.5 minutes.13
Fig. 5 shows the median average radius for the “standard” and attentional field tests as a function of age group. The error bars show the 25th and 75th percentiles. The values for the “standard” condition show little loss with age as expected for these suprathreshold targets, but the loss of visual field with attentional load is dramatic. For the oldest age group (90+), the median radius is only 25% of maximum. In addition, 25% of this age group did not see any of the peripheral targets when attending to the central task (% of maximum near zero for 75th percentile) even though the radius in “standard” conditions was 88% of maximum. The attentional task we used is fairly simple and poses no difficulties for the youngest of the older observers, who performed similarly on the standard and attentional tasks. This group (<60 years) performed no differently than young adults on these tasks. The fall off with age for the attentional task is substantial and starts before the age of 70. In the oldest age group, any attentional load essentially makes suprathreshold targets disappear, which may contribute to driving difficulties because the attentional visual field has been shown to be correlated with accidents.20 Clearly, the attentional visual field measure consists of a visual task and a cognitive component. Our sample is well educated (93% finished high school and 50% finished college) and by and large cognitively sound, not frail or demented. We assessed the cognitive status of the participants using the Short Portable Mental Status Questionnaire (SPMSQ).21 Only 0.08% of the original 900 SKI study participants were considered “moderately to severely intellectually impaired” (a score of ≥5 on the SPMSQ21), whereas the rest (99.2%) were classified as having intact cognition or only mildly impaired. Thus, even subtle changes in cognitive status contribute to the performance on the attentional visual field test.
Conclusions from Cross-sectional Vision Data
The results from the cross-sectional vision data have demonstrated that each vision function has its own “aging function” even though the spatial vision measures have the same exponential shape. Unlike most other vision measures, standard high contrast visual acuity shows mild changes with age, even among the very old. In fact, only CFF shows less change with age than high contrast visual acuity. Many elderly persons have significant visual impairment on other spatial vision measures despite having good acuity. Not surprisingly given the increased scatter in the eye and the high prevalence of cataracts, disability glare is the vision function most affected by aging. The dynamics of adaptation are also slowed as evidenced by the glare recovery results. Gross stereopsis and attentional visual fields are reduced, and color vision changes occur as expected from the yellowing of the lens. The changes in the nonstandard vision functions cannot be predicted based on visual acuity on an individual basis even though the measures are correlated.17 To appreciate the vision function of elderly persons, it is imperative to use other measures in addition to visual acuity.
We tested three measures of visual performance: reading for unrelated letters and words using the Pepper Visual Skills for Reading Test, visual search for grocery items, and face recognition. The reading results for those with good visual acuity have been published.22 Surprisingly, even those with 20/32 or better visual acuity showed significant losses in reading performance. The accuracy of reading unrelated words out loud was maintained, but the rate of reading slowed down considerably. The corrected reading rate for those <65 years was well over 100 words/min, whereas the rate was 70 words/min for the oldest age group (85+ years) despite the fact that all the observers had at least 20/32 visual acuity. When performance on low contrast vision tests, motor ability, and attentional field integrity were taken into account, age was not a significant predictor of reading rate. There was no relation between performance on the test of cognitive status (SPMSQ) and reading when vision variables were included in the model. This may only reflect the insensitivity of the SPMSQ to subtle changes in cognitive status because the attentional visual field (which has a cognitive component) was significantly related to reading performance. Our results strongly suggest that reading performance should be assessed in older persons, even when acuity is good.
The relation between vision function measures and general health and functioning has also been published.14 Poor performance on many of the vision function measures was significantly associated with particular measures of physical performance. For example, poor visual field integrity and poor adaptation were associated with mobility limitations, and poor binocularity was associated with failure on a chair-stand task. Spatial vision was not a significant correlate of physical performance when other vision functions were included in the model.
Another measure with surprisingly poor performance in old age is face recognition.23 Face recognition was measured at the second visit using slides of four faces of different sizes with four different expressions. The sizes of the test faces were equivalent to an adult face presented at distances from 0.75 to 24 m. The same four faces with neutral expressions were present and visible at all times at a close distance. The neutral faces of two women and two men were named. The task of the observer was to identify the person on the slide by name. The expression on the face also had to be identified (sad, happy, angry, or afraid). The slides were shown for 10 s. The threshold was determined using probit analysis. The results are presented in equivalent viewing distance (EVD); that is, what is the furthest distance that the observer can correctly identify a normally sized adult face.
Fig. 6 shows the mean EVD in meters for identification and expression as a function of age group for 576 older observers. With increasing age, the EVD became much shorter, with those >85 years needing as short a distance as 2.1 m to correctly name the person in the slide and their expression. Standard multiple regression analyses revealed that age, gender, years of education, cognitive status, spatial vision function, and stereoacuity were significant independent predictors of “Person and Expression” face recognition. In this study, the Mental Alternation Test (MAT) was used as the cognitive measure.24 The MAT accounted for 1% of the independent variance in the model.
SKI Study: Longitudinal Results
Five hundred ninety-six of the 900 original participants were retested on average 4.4 years later (SD, 1.0 year). Of the 596 retested, 90% had visual acuity of 20/40 or better at baseline (N = 537). A substantial number of those with initially relatively good acuity (better than 20/40) showed a subsequent visual acuity loss with a rate of three or more lines (doubling of threshold) per decade. Can the other baseline vision measures help identify those persons who are at risk of subsequent high contrast visual acuity loss? To answer this question, information about ocular health status and refractive error is necessary. Medical eye records were requested from the eye care professional of each participant. Those without medical records were removed from this analysis. In addition, if the records indicated that changes in refractive error could have been responsible for the loss of acuity, that person was excluded. Four hundred twenty-six participants with medical eye records and acuity 20/40 or better at baseline were included in this analysis. Of these, 16.9% subsequently lost visual acuity of 0.3 log units/decade or more (three lines or more/decade).
Fig. 7 shows the percentage of those who subsequently lost 0.3 log units/decade of high contrast acuity as a function of low contrast acuity in glare measured at baseline. Clearly, the worse acuity in glare at baseline, the higher percentage of observers who subsequently lost high contrast visual acuity. For those with poor acuity in glare (>1.1 logMAR or worse than 20/253), 45.2% subsequently lost visual acuity, whereas for those with the best acuity in glare (<0.5 logMAR or better than 20/63), only 2.6% subsequently lost visual acuity. In a multiple logistic regression analysis including all the vision functions, age, gender, retinal disease, and cataract status, only low contrast vision function was a significant predictor of future acuity loss. Age and retinal disease status, which were predictors in univariate analyses, dropped out of the model when low contrast vision function was included. High contrast acuity at baseline was not a significant predictor of future acuity loss. For each doubling of low contrast threshold at baseline (0.3 log units), the person is 2.35 times more likely to subsequently lose visual acuity. Each of the low contrast vision measures of contrast sensitivity, low contrast low luminance acuity (SKILL Card), or acuity in glare produced similar results. These results were recently published.25
Aging studies or clinical evaluations relying solely on standard high contrast visual acuity to measure vision function grossly underestimate the true age-related loss in vision function. Even persons with good standard acuity are visually impaired in conditions of reduced contrast, reduced lighting, changing light level, or reduced contrast in the presence of glare. Among elderly persons with initially relatively good acuity, deficits in certain other measures of vision function are significant predictors of subsequent loss of standard visual acuity. Elderly persons with good visual acuity who score at the low end of the range on low contrast low luminance acuity or contrast sensitivity or acuity in glare have a nearly 50% chance of significant visual acuity loss in the next few years. Thus, the low contrast tests detect subtle subclinical ocular changes that the standard visual acuity measures are unable to detect.
The predictive power of the low contrast tests has implication for clinical trials of preventative vision therapies because the persons at highest risk can be identified by measuring low contrast vision function at baseline, making trials more efficient by focusing the preventative therapies only to those at high risk who are likely to develop future vision loss.
In older patients with relatively normal high contrast visual acuity, performing simple vision tests using low contrast low luminance targets or disability glare or contrast sensitivity tests may also help practitioners identify and closely monitor those at high risk for subsequent vision problems.25
Supported by National Eye Institute research grant EY09588 to John Brabyn and the Smith-Kettlewell Eye Research Institute.
School of Optometry
University of California at Berkeley
Berkeley, CA 94720-2020
2.Kahn HA, Leibowitz HM, Ganley JP, Kini MM, Colton T, Nickerson RS, Dawber TR. The Framingham Eye Study. I. Outline and major prevalence findings. Am J Epidemiol 1977;106:17–32.
3.Tielsch JM, Sommer A, Witt K, Katz J, Royall RM. Blindness and visual impairment in an American urban population. The Baltimore Eye Survey. Arch Ophthalmol 1990;108:286–90.
4.Klein R, Klein BE, Linton KL, De Mets DL. The Beaver Dam Eye Study: visual acuity. Ophthalmology 1991;98:1310–5.
5.Rubin GS, West SK, Munoz B, Bandeen-Roche K, Zeger S, Schein O, Fried LP. A comprehensive assessment of visual impairment in a population of older Americans. The SEE Study. Salisbury Eye Evaluation Project. Invest Ophthalmol Vis Sci 1997;38:557–68.
6.Owsley C, Sekuler R, Siemsen D. Contrast sensitivity throughout adulthood. Vision Res 1983;23:689–99.
7.Bailey IL, Bullimore MA. A new test for the evaluation of disability glare. Optom Vis Sci 1991;68:911–7.
8.Winn B, Whitaker D, Elliott DB, Phillips NJ. Factors affecting light-adapted pupil size in normal human subjects. Invest Ophthalmol Vis Sci 1994;35:1132–7.
9.Pokorny J, Smith V, Lutze M. Aging of the human lens. Appl Opt 1987;26:1437–40.
10.van den Berg TJ. Analysis of intraocular straylight, especially in relation to age. Optom Vis Sci 1995;72:52–9.
11.Siik S, Airaksinen PJ, Tuulonen A. Light scatter in aging and cataractous human lens. Acta Ophthalmol (Copenh) 1992;70:383–8.
12.Elliott DB, Whitaker D. Changes in macular function throughout adulthood. Doc Ophthalmol 1990;76:251–9.
13.Haegerstrom-Portnoy G, Schneck ME, Brabyn JA. Seeing into old age: vision function beyond acuity. Optom Vis Sci 1999;76:141–58.
14.West CG, Gildengorin G, Haegerstrom-Portnoy G, Schneck ME, Lott L, Brabyn JA. Is vision function related to physical functional ability in older adults? J Am Geriatr Soc 2002;50:136–45.
15.West CG, Gildengorin G, Haegerstrom-Portnoy G, Lott LA, Schneck ME, Brabyn JA. Vision and driving self-restriction in older adults. J Am Geriatr Soc 2003;51:1348–55.
16.Schneck ME, Haegerstrom-Portnoy G. Practical assessment of vision in the elderly. Ophthalmol Clin North Am 2003;16:269–87.
17.Haegerstrom-Portnoy G, Schneck ME, Lott LA, Brabyn JA. The relation between visual acuity and other spatial vision measures. Optom Vis Sci 2000;77:653–62.
18.Brabyn JA, Schneck M, Haegerstrom-Portnoy G, Lott L. Functional vision: “Real World” Impairment, Examples from the SKI Study. Visual Impair Res 2004;6:35–44.
19.Schneck ME, Haegerstrom-Portnoy G, Lott LA, Brabyn JA. Ocular contributions to age-related loss in coarse stereopsis. Optom Vis Sci 2000;77:531–6.
20.Owsley C, Ball K, McGwin G, Sloane ME, Roenker DL, White MF, Overley ET. Visual processing impairment and risk of motor vehicle crash among older adults. JAMA 1998;279:1083–8.
21.Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc 1975;23:433–41.
22.Lott LA, Schneck ME, Haegerstrom-Portnoy G, Brabyn JA, Gildengorin GL, West CG. Reading performance in older adults with good acuity. Optom Vis Sci 2001;78:316–24.
23.Lott L, Haegerstrom-Portnoy G, Schneck ME, Brabyn JA. Face and object recognition in elders. Invest Ophthalmol Vis Sci 2001;42:S863.
24.Billick SB, Siedenburg E, Burgert W, 3rd, Bruni-Solhkhah SM. Validation of the Mental Alternation Test with the Mini-Mental State Examination in geriatric psychiatric inpatients and normal controls. Compr Psychiatry 2001;42:202–5.
25.Schneck ME, Haegerstrom-Portnoy G, Lott LA, Brabyn JA, Gildengorin G. Low contrast vision function predicts subsequent acuity loss in an aged population: the SKI study. Vision Res 2004;44:2317–25.