Postural Stability in the Elderly during Sensory Perturbations and Dual Tasking: The Influence of Refractive Blur

Abstract

purpose. To determine the influence of refractive blur on postural stability during somatosensory and vestibular system perturbation and dual tasking.

methods. Fifteen healthy, elderly subjects (mean age, 71 ± 5 years), who had no history of falls and had normal vision, were recruited. Postural stability during standing was assessed using a force platform, and was determined as the root mean square (RMS) of the center of pressure (COP) signal in the anterior-posterior (A-P) and medial-lateral directions collected over a 30-second period. Data were collected under normal standing conditions and with somatosensory and vestibular system perturbations. Measurements were repeated with an additional physical and/or cognitive task. Postural stability was measured under conditions of binocular refractive blur of 0, 1, 2, 4, and 8 D and with eyes closed. The data were analyzed with a population-averaged linear model.

results. The greatest increases in postural instability were due to disruptions of the somatosensory and vestibular systems. Increasing refractive blur caused increasing postural instability, and its effect was greater when the input from the other sensory systems was disrupted. Performing an additional cognitive and physical task increased A-P RMS COP further. All these detrimental effects on postural stability were cumulative.

conclusions. The findings highlight the multifactorial nature of postural stability and indicate why the elderly, many of whom have poor vision and musculoskeletal and central nervous system degeneration, are at greater risk of falling. The findings also highlight that standing instability in both normal and perturbed conditions was significantly increased with refractive blur. Correcting visual impairment caused by uncorrected refractive error could be a useful intervention strategy to help prevent falls and fall-related injuries in the elderly.

Falls are a major cause of death and morbidity in older people. Of the 7000 people who died because of falls in England and Wales during 1994 and 1995, 84% were aged more than 65 years. ¹ In the United States, approximately 10,000 deaths each year are related to falls among elderly patients. ² In attempting to determine the etiology of falls in the elderly, investigators in many studies have shown an association between postural stability during standing and increased risk of falling, ^{3 4 5} and many have shown the visual contribution to postural stability to be greater in fallers than in nonfallers. ^{6 7}  

The role of visual information in balance control is well documented. In 1946 Edwards ⁸ demonstrated that the availability of visual information could reduce postural instability by as much as 50%. Vision is particularly important in regulating balance under more challenging conditions. For example, under conditions of reduced somatosensory input, postural stability has been shown to correlate with measures of visual acuity (VA), ^{9 10} contrast sensitivity (CS), ^{6 10 11} and stereoacuity. ¹⁰ Similarly, vestibular system perturbations have a pronounced effect on postural stability when vision is also disrupted. ^{12 13 14 15} This suggests that poor vision is an important factor in determining risk of falling, and prevalence studies support this notion. For example, Jack et al. ¹⁶ reported that in a group of 200 elderly patients admitted to an acute geriatric clinic, 101 (50.5%) had impaired vision (best eye acuity worse than 20/60 Snellen). More particularly, a high prevalence (76%) of visual impairment was found in the patients admitted because of falls, and the authors highlighted that 79% of the visual impairment was reversible, mainly by correcting refractive errors (40%). 

Although correcting refractive errors is the most common treatment of visual impairment, relatively few studies have been undertaken to determine the effect on postural stability of refractive blur, and most of these have involved young subjects. ^{8 17 18 19} Edwards ⁸ found an increase in median body sway of approximately 51% (mean increase 28%) with the addition of a +5-D lens in 50 young adults. Paulus et al. ¹⁸ found a 25% increase in postural sway when five young subjects with myopia removed their spectacles (four subjects had myopic errors between −3 and −5 D and one had a myopic correction of −11 D). They also found, in a later study, a 25% increase in postural sway with +4- and +6-D blur in a group of 10 young subjects, which increased to 50% and nearly 100%, respectively, with +8- and +10-D blur. ¹⁷ However, in a third study, they found a much smaller (∼10%) and nonsignificant increase in postural sway when 16 subjects with myopia and hyperopia between 2 and 5 D removed their spectacles (the ages of the subjects was not reported). ¹⁸ In a previous study in young subjects, we found a significant increase in postural instability with refractive blur that was greater with perturbation of the somatosensory and/or vestibular systems. ¹⁹  

Most of these studies highlight the importance of reducing refractive blur to maintain postural stability, particularly when the somatosensory input has been perturbed. However, no study has been undertaken to determine the effects of refractive blur during multisensory perturbations—that is, when somatosensory and/or vestibular inputs and/or cognitive function are disrupted. In addition, only one study has examined the effects of refractive blur on postural stability in the elderly. ¹⁷ This study reported that five elderly patient with aphakia and one elderly subject with high myopia showed no difference in postural sway with and without their spectacles. Elderly subjects have been shown to have the greatest risk of falling ^{1 2} and are most at risk of having out-of-date spectacles that cause visual impairment. ^{16 20 21} Thus, the purpose of the present study was to determine the effects of binocular refractive blur on the postural stability of elderly subjects, when input from both the somatosensory and vestibular systems were perturbed and when performing additional cognitive and everyday physical tasks. 

Methods

Subjects

Fifteen elderly subjects, nine women and six men (mean ± SD age, 71.06 ± 5.40 years) were recruited from a group of volunteer patients who attend the University of Bradford Eye Clinic for teaching purposes. The tenets of the Declaration of Helsinki were observed, and the study gained approval from the university ethics committee. Informed consent was obtained after the nature of the study had been fully explained. An assessment was made to ensure that all subjects had no history of falls. For this study, a fall was defined as falling all the way to the floor or ground, falling and hitting an object such as a chair or stair, or falling from one level to another, for example from bed to the ground. ²² Subjects were also screened using a self-report health questionnaire. Subjects with any cardiac arrhythmias, vestibular disturbances, diabetes, or severe arthritic conditions and medications affecting balance were excluded. Scores on the Lawton Activities of Daily Living (ADL) questionnaire ²³ were high, with all subjects scoring above 14 and 12 subjects scoring a maximum 16. This indicated that the subjects were all independently mobile. To ensure that any changes to postural stability due to head extensions were caused by disruption of the input from the vestibular system and not vertebrobasilar insufficiency ¹⁵ and to ensure that the subjects would be able to participate in the study safely, they each underwent a preliminary and supplementary vertebral artery examination. ²⁴ None of the subjects had any complications such as dizziness, diplopia, fainting, dysarthria, and dysphasia after the various head extensions and rotations involved in the examination. 

Measurements of VA and ocular screening using slit lamp biomicroscopy, tonometry, indirect ophthalmoscopy, and central visual field screening were undertaken. To ensure that vision loss was entirely due to refractive error, subjects with a history of amblyopia, strabismus, eye disease or ocular surgery, binocular VA less than 0.0 logarithm of the minimum angle of resolution (logMAR; Snellen equivalent, ∼20/20) and/or any visible ocular disease were excluded. Subjective refraction was performed to obtain the subject’s optimal refractive correction at 4 m. Binocular visual function was subsequently assessed by measuring VA and CS. Binocular VA was measured (mean VA −0.06 ± 0.04 logMAR, Snellen equivalent, ∼20/15) using the Early-Treatment Diabetic Retinopathy Study (ETDRS) logMAR chart, with a letter-by-letter scoring system and chart luminance of 160 cd/m² and a 4-m working distance. Binocular CS was measured (mean, 1.60 ± 0.11 log CS) using the Pelli-Robson chart at 1 m with a letter-by-letter scoring system and a chart luminance of 200 cd/m². Binocular VA and CS were subsequently remeasured with application of additional binocular blur lenses of +1, +2, +4, and +8 spheric diopter (DS) in a randomized order. 

Postural Stability Measures

Postural stability measurements were determined while subjects stood stationary on a force platform (model OR6-7; Advanced Medical Technology Inc. Boston, MA) mounted so its surface was flush with the floor. Displacements of the center of pressure (COP) in the anterior-posterior (A-P) and medial-lateral (M-L) directions were derived from the force and moment profiles measured by the force plate. Fluctuations in the displacement of the COP signal were quantified using the root mean square (RMS) of the amplitude. These fluctuations reflect the response of the CNS to changes of the center of mass. ^{25 26} The subjects were asked to stand still on the force plate for 30-second periods with their arms by their sides and their feet placed so that the inner edges of both feet were one foot-length (their own) apart. They were asked to fixate the middle of a visual target that consisted of a horizontal and vertical square-wave pattern with a fundamental spatial frequency of 2.5 cyc/deg. ^{13 27} Intermediate spatial frequencies have been shown to provide better visual stabilization of posture than lower or higher frequencies. ²⁶ The pattern had a Weber contrast of 25%, which we assumed to be representative of contrast levels typically found in a home environment, and the target covered an area of 1.1 m² with a viewing distance of 1 m. The target was adjusted for the height of each subject so that its center was at eye level. Viewing was binocular, and each subject’s vision was corrected with the optimal 4-m refractive correction and a 0.75-DS working-distance lens with full-aperture lenses in a trial frame at a distance of 1 m. 

Standing postural stability was measured under the following conditions:

1.  
   Normal standing. Normal, quiet standing on a solid surface (bare platform).
2.  
   Normal standing with cognitive and physical tasks. Normal, quiet standing on a solid surface (bare platform) while performing a cognitive and everyday physical task. The physical task involved the subjects holding a wooden tray (54 × 35 cm) that contained three (empty) foam cups. For the cognitive task, subjects were asked to count backward in a series of twos from a designated number that was randomly chosen by the investigator.
3.  
   Somatosensory input disrupted. Quiet standing on a 16-cm foam surface. The compliant nature of the foam makes it difficult for the kinesthetic system to provide accurate body orientation information in relation to the ground and thus disrupts somatosensory system inputs. ^{7 10}
4.  
   Somatosensory input disrupted with physical task. Standing on the foam surface while performing the everyday physical task.
5.  
   Somatosensory input disrupted with cognitive and physical tasks. Standing on the foam surface while performing the cognitive and everyday physical tasks.
6.  
   Somatosesory and vestibular input disrupted. Standing on the foam surface with head hyperextended to 45°. ^{13 15} The head extension places the utricular otoliths beyond their normal working range, causing a disruption of the vestibular system input. ¹⁵ To ensure that there was no change in the visual input for this condition, the visual target was raised and oriented to a 45° position to ensure the same visual target and test distance as in the control condition.

Under each of the six test conditions, standing balance was measured with the optimal refractive correction for the 1-m working distance; under binocular blur levels of +1, +2, +4, and +8 D; and with eyes closed. Before testing began, subjects participated in a familiarization session, in which they stood on a foam surface so that they could become familiar with standing with somatosensory input disrupted. Subjects were also exposed to the various blur conditions to prepare them for the visual conditions. The order of the 36 postural stability measurements was randomized. After each 30-second trial, subjects were allowed a rest period of 1 minute, when they could be seated. 

COP RMS data were analyzed with the generalized estimating equation (GEE) population-averaged model in a statistical-analysis program (Stata ver. 7.0 statistical program; Stata Corp., College Station, TX), which accounted for the correlation of readings within subjects. An exchangeable correlation structure was judged to be appropriate, given the experimental design. Level of significance was set at P < 0.05. The terms in the model were:

1.  
   A fixed factor with two levels: A-P and M-L directions of stability.
2.  
   Sensory disruption: a fixed factor with the six levels described earlier.
3.  
   Blur: a fixed factor with six levels: eyes open with no blur; 1-, 2-, 4-, and 8-D blur and eyes closed. The eyes-closed condition was not included in this factor. However, it was measured for each sensory disruption condition and allowed subsequent comparison of eye-closed measurements with those with eyes open and various amounts of blur.

Results

The effect of refractive blur on VA and CS is shown in Figure 1 . The GEE population-averaged model of the COP RMS data were checked by plotting the predicted values of stability against the actual values and against the Studentized residuals. There was generally close agreement between the actual and predicted values of stability, and departures from model assumptions were not severe, which suggests that the model is a good approximation of the data. The interactions of blur and sensory disruption, and blur and APML were found to be statistically significant and were therefore included in the model. Given these significant interaction terms, subsequent testing of compound hypotheses (such as the effects of blur and sensory disruption) involved comparing conditions simultaneously over all relevant interaction terms using the “accumulate” option in the software (Stata; Stata Corp.). In this way, the effect between conditions was assessed by incorporating the relevant main conditions and interaction terms. 

The changes in the COP RMS in the A-P and M-L directions as a function of refractive blur for both the normal standing condition and with somatosensory perturbation are shown in Figure 2 . The APML COP RMS term in the model was highly significant (χ² = 90.8, P < 0.001), indicating that there were highly significant differences between postural stability in the A-P and M-L directions in all conditions. The mean COP RMS displacement in the A-P direction was always greater than in the associated M-L direction (Fig. 2) . The APML COP RMS-blur interaction term in the model was also significant (χ² = 25.7, P < 0.0001), indicating that blur had a different affect on A-P stability than it did on M-L stability. The data also indicate that the effects of somatosensory perturbation and refractive blur were similar for both A-P and M-L directions (Fig. 2) . 

To avoid being repetitive and because the effects on postural instability were greater in the A-P direction, all results and analyses will subsequently point out changes in A-P COP RMS. A-P COP RMS measurements as a function of refractive blur and for the six measurement conditions are shown in Table 1 and Figure 3 . Sensory disruption, blur, and their interaction were highly significant factors (P < 0.01) in the GEE population-averaged model. Given the significant interaction term, subsequent testing of the effects of blur and sensory disruption involved comparing relevant conditions simultaneously over all relevant interaction terms so that the main effects were compared while incorporating the interaction term into the compound hypothesis tested. Wald χ² ₄ results are presented in Table 2 and show the significance of the differences in A-P COP RMS between the main stance conditions. The degrees of freedom for the χ² statistics relate to the number of hypotheses (determined by the number of combinations of appropriate interaction terms) conjointly tested. There were significant differences in A-P COP RMS between all measurement conditions, except between standing with somatosensory disruption and standing with somatosensory disruption while performing the physical everyday task (P = 0.31; Table 2 ). 

Refractive blur was shown to have a significant effect on postural stability under normal standing conditions (χ² ₄ = 22.24, P = 0.0002). Table 3 indicates whether the effect of refractive blur on postural stability was significantly different from one standing condition to another. Once again, these effects were compared while incorporating the interaction terms into the compound hypothesis tested. 

Discussion

Blur and Visual Function

Blur had a significantly greater effect on logMAR VA than on Pelli-Robson CS at low levels of blur (Fig. 1) , which agrees with previous findings. ²⁸ The Pelli-Robson chart measures CS at or slightly below the peak of the CS function at approximately 1.5 cyc/deg, and these spatial frequencies are known to be relatively unaffected by small amounts of refractive blur. ²⁸  

M-L Versus A-P Postural Stability

The mean COP RMS displacement in the A-P direction was always greater than the associated M-L measure (Fig. 2) , which is in agreement with previous findings. ^{25 26} The APML COP RMS–blur interaction term in the model was significant, indicating that blur had a different effect on A-P stability than it did on M-L stability. This agrees with previous findings ^{19 29} and suggests that refractive blur may have a greater effect on the visual stimuli that provide information to control A-P stability than on the stimuli providing information for M-L stability. 

Effect of Sensory Perturbation

Table 1 and Figure 3 indicate the changes in A-P COP RMS for the different standing conditions. The differences in A-P COP RMS between the sensory disruption conditions were all significant with the exception of the addition of the physical task (Table 2) . It is well known that a disruption of the input from the somatosensory system increases postural instability and that this places greater emphasis on the visual system. ^{6 10 11 30 31 32} For example, several studies have reported significant differences in postural sway in groups of visually impaired subjects, compared with visually normal subjects when somatosensory input was disrupted. ^{6 10 11 31 32} Similarly, Lord et al. ^{7 31} and Lord and Menz ¹⁰ reported that increases in postural sway during standing with somatosensory perturbations were mainly associated with poor VA and CS. Other studies have reported increases in postural sway with eyes closed compared with eyes open under conditions of somatosensory system disruption. ^{12 33 34 35} These findings, which are corroborated by findings from the present study, indicate that the visual system plays a greater role in maintaining postural stability when somatosensory information is disrupted. This is emphasized by findings from the present study, indicating that although refractive blur increased postural instability under normal standing conditions, the increase was much larger when somatosensory input was disrupted (Fig. 3) . 

In investigating the role of the vestibular system in postural control, previous researchers have perturbed stability by asking subjects to stand with the head hyperextended by 45°. ^{12 13 15} Although these studies indicate that such perturbations have a significant effect on postural instability, the visual target used in the different testing condition was not controlled, ^{12 13 15} —that is, subjects looked at a visual target during normal standing and at the uniform white ceiling during head extension. Thus, the increased postural instability may have resulted from a change in the visual information available. In the present study, the visual target was maintained during hyperextension of the head, and therefore the postural instability increases can be directly attributed to the disruption of the vestibular system. This indicates that physiological disruptions to the vestibular system input caused by head hyperextensions (such as looking upward when negotiating stairs, looking to a cupboard above eye level or when changing a light bulb) can increase postural instability in elderly subjects, particularly under conditions of reduced somatosensory input. It would be of great interest to determine whether head flexion (such as when walking downstairs or stepping over obstacles, particularly when wearing multifocal lenses) causes changes to postural stability similar to those caused by hyperextending the head; we are currently testing this hypothesis. 

Influence of Refractive Blur

In the present study the effect of refractive blur was measured under normal standing conditions and when somatosensory and vestibular inputs and cognitive function were all disrupted. Findings indicate that blur significantly increased postural instability under normal, quiet standing conditions (Fig. 3) . Furthermore, refractive blur increased postural instability to an even greater extent when disruption to the inputs from the somatosensory system and disruptions to both the somatosensory and vestibular systems were present (Fig. 3 ; Tables 1 3 ). The effect of refractive blur may depend on the visual target used, ¹⁹ and this may explain slight discrepancies when comparing the findings of previous studies. The present study and an earlier one from our laboratory involving young subjects ¹⁹ suggest a larger effect of refractive blur on postural stability than the studies of Straube et al. ¹⁷ and Paulus et al. ^{18 29} Given that the effect of dioptric blur was influenced by the usefulness of the input from the somatosensory and vestibular systems (Fig. 3) , these differences are probably due to methodological differences in how the experimental setup disrupted the input from these systems. In addition, it is possible that the visual target used by Straube et al. ¹⁷ and Paulus et al. ^{18 29} (a screen randomly covered with different colored dots of different sizes) did not provide enough visual information to aid postural stability, so that dioptric blur of the target had less chance to disrupt stability. 

The refractive blur used in the present study caused an acute change in the visual information provided to postural stability. The effect is similar to a patient’s removing his or her spectacles. It is possible that chronic changes in refractive error (such as myopia induced by nuclear cataract) or other chronic changes to vision caused by age-related ocular disease, such as cataract or age-related macular degeneration, ^{11 32} have less of an effect on postural stability because compensation mechanisms may have been acquired over time. 

Cognitive and Physical Dual Tasks

In the present study, the effect of blur during somatosensory disruptions did not change when the cognitive and everyday physical dual tasks were added (Table 3) . We also found that the physical task alone had no significant influence on postural stability (Table 2) , which suggests that once the subjects had made the necessary postural adjustments when holding the tray, the adjustments could be maintained and did not require any further modifications. In comparison, the cognitive task (in conjunction with the physical task) showed a significant effect, across all blur conditions, on postural stability (Table 2) , which may have been due to its being a continuous task requiring attention throughout the trial session. Investigators who have assessed the cognitive influences on postural stability in previous studies have stated similar findings. Shumway-Cook et al. ³⁶ and Melzer et al. ³⁷ showed increases in postural instability with the addition of a cognitive task. However, the tasks the subjects were asked to complete were presented visually and hence would require not only cognitive functioning but also visual functioning. In particular, the optokinetic movements required in reading the visual display may have partially contributed to the increased postural instability observed. In a later study Shumway-Cook and Woollacott ³⁸ used an auditory cognitive task, and in a study by Condron and Hill, ³⁹ subjects were asked to count backward. Both investigations found an increase in postural instability (under reduced somatosensory input conditions), when subjects were asked to complete the cognitive task. Performing a concurrent cognitive task has been shown to affect postural instability more in the elderly than in the young ^{39 40} and has also been shown to have a greater effect in subjects with a history of falls. ^{36 38 39} The study by Rankin et al., ⁴⁰ found reduced muscular activity in both the gastrocnemius and tibialis anterior when subjects were asked to perform a cognitive task. This effect was suggested to be because less attention was available for balance control. In the present study, the lack of a blur effect for the cognitive task (despite its effect of increasing postural instability across blur conditions) suggests that cognitive functioning does not act on postural stability in the same way as the somatosensory and vestibular systems, and this suggests that it occurs at a different stage in the mechanism controlling postural stability. Furthermore, it is possible that the effects of the cognitive task may have been greater if a more difficult cognitive task had been used. 

Summary

Findings indicate that increasing levels of refractive blur can significantly increase postural instability in the elderly, particularly when the information provided by the somatosensory and/or vestibular systems is disrupted. Poor quality input from these sensory systems may occur with thick carpeting or shoes, ^{41 42} when negotiating stairs, when looking or reaching to a cupboard above eye level, ^{12 14} or in the presence of various systemic diseases. ^{12 43 44} Our findings also indicate that when elderly subjects perform a cognitive dual task, postural instability can increase further. The findings suggest that the factors that cause increases in postural instability (disruption to the somatosensory, visual and vestibular sensory systems, and dual cognitive tasking) have a cumulative effect, which indicates why the elderly, many of whom have poor vision and musculoskeletal and CNS degeneration, are at greater risk of falling. Findings also highlight that standing stability in both normal and perturbed conditions was improved with reduced refractive blur. Correcting visual impairment in the elderly caused by uncorrected refractive error ^{16 20 21} could thus be an intervention strategy to help prevent falls and fall-related injuries. 

 Supported by Grant 3991/882 from the PPP Foundation and by the Vision Research Trust.

 Submitted for publication October 7, 2002; revised December 12, 2002; accepted January 3, 2003.

 Disclosure: V. Anand, None; J.G. Buckley, None; A. Scally, None; D.B. Elliott, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Mean logMAR visual acuity and Pelli-Robson CS scores as a function of refractive blur.

Figure 2.

 

Mean ± SE COP RMS measurements in the A-P and M-L directions as a function of refractive blur. COP RMS data are shown for normal standing conditions and under somatosensory system perturbation.

Table 1.

 

Mean A-P COP RMS during All Standing Conditions for the Eyes-Open with Optimal Correction 0-D and 4-D Blur Conditions

Sensory Condition	COP 0 D (mm)	% Increase from Normal Standing	COP +4 D (mm)	% Increase from Normal Standing	% Increase from 0-D to 4-D Blur
Normal standing	3.3	0	4.0	0	23
Normal standing with cognitive and physical task	4.0	24	5.1	28	26
Somatosensory perturbation	6.3	93	8.4	109	33
Somatosensory perturbation with a physical task	7.0	114	9.4	133	34
Somatosensory perturbation with physical and cognitive tasks	7.2	121	9.8	143	35
Somatosensory and vestibular system perturbation	8.2	150	11.5	186	41

Figure 3.

 

Mean ± SE COP RMS measurements in the A-P direction as a function of refractive blur. Mean A-P COP RMS data are shown for subjects under all sensory conditions.

Table 2.

 

Significance of Changes in A-P COP RMS between Standing Conditions across All Blur Conditions

Normal standing vs. normal standing with cognitive and physical task	χ2 4 = 26.05; P < 0.0001
Normal standing vs. standing with somatosensory perturbation	χ2 4 = 39.6; P < 0.0001
Standing with somatosensory perturbation, with and without a physical task	χ2 4 = 4.8; P = 0.31
Standing with somatosensory perturbation, with and without a physical and cognitive task	χ2 4 = 16.4; P = 0.0025
Standing with somatosensory perturbation vs. standing with somatosensory and vestibular system perturbation	χ2 4 = 130.0; P < 0.0001

Table 3.

 

Wald χ² and Probabilities Indicating Whether the Effect of Refractive Blur on Postural Stability Was Significantly Different between Pertinent Standing Conditions

Conditions	χ2 and P
Normal standing and normal standing with a cognitive and physical task	χ2 3 = 1.7 P = 0.63*
Normal standing and standing with somatosensory perturbation	χ2 3 = 22.2 P = 0.0002
Standing with somatosensory perturbation, with and without a physical task	χ2 3 = 3.1 P = 0.38*
Standing with somatosensory perturbation, with and without a physical and cognitive task	χ2 3 = 0.5 P = 0.91*
Standing with somatosensory perturbation and standing with somatosensory and vestibular system perturbation	χ2 3 = 8.9 P = 0.031