Aging Clin . Exp . Res. 4: 219-225,1992
Changes in posture and balance with age A.P. Woodhull-McNeal School of Natural Science, Hampshire College, Amherst, Massachusetts, U.S.A. ABSTRACT. Changes in posture with age are of concern because of their association with impaired mobility and the possibility of falls. In this study alignment of the joint centers and balance of body segments were measured in 41 women , aged 65 and older, and correlated with weight, age, change in height, and activity. Balance was assessed by calculating the whole-body center of gravity, and the partial centers of gravity above the knee and hip. Compared to a younger reference population, the older subjects had greater kyphosis , a more posterior hip position , and leaned forward more from the hips (more anterior center of gravity above the hips). Among the 41 elderly subjects, larger kyphoses correlated with greater decrease in height. Elderly subjects who were inactive tended to lean more, but forward lean was not correlated to other variables. These findings suggest two separate sets of changes: an osteoporosis-related stooping, and also a forward lean that may relate to muscular weakness or to fear of instability. Posture among the elderly was quite varied and did not follow anyone pattern of change. (Aging Clin . Exp. Res. 4: 219-225, 1992) INTRODUCTION Among the numerous changes associated with aging, changes in posture seem to signal major changes in function. The stereotypical picture of old age is one of a stooped-over torso with exaggerated kyphosis, forward-leaning upper body, and bent knees . Osteoporosis can cause some of these changes, since compression fractures lead
to "wedging" of the anterior side of the vertebrae, causing both increased kyphosis and decreased lumbar lordosis (1); the lumbar changes may in turn lead to changes in pelvic tilt and knee position . Arthritis, muscle weakness, and other conditions may also contribute to different postural changes. Most studies concerning posture in the elderly have focussed on postural sway. (2-5). This is a logical concern since increased sway may be a sign of generalized instability, and thus a sign of increased danger of falling. Not surprisingly , sway has been found to increase with age (4, 6), and old people who fall have on average increased postural sway (5, 7). Changes in gait are another major focus of research (8). It was suggested that increased forward leaning of head , arms and trunk (HAT) is part of a complex of changes in gait associated with instability (9). Standing posture, including the static balance of body segments, has been studied very little, although it could be important in understanding some of the reasons for postural instability. When posture changes , then the biomechanics of standing and of locomotion must also change, which could add to mechanical instability and insecurity. Brocklehurst et al. (10) correlated some measures of skeletal change to postural sway, and found that increased sway was associated with larger loss of height, greater knee flexion, and less grip strength. In young adults the skeleton is aligned for maximal stability in several ways. Although not strictly linear, the joint centers are aligned fairly linearly above and forward of the ankle joint (11 , 12) . The center of gravity of the whole
Key words : Aging, balance, posture. Correspondence: Ann P. Woodhull-McNeal, PhD., School of Natural Science, Hampshire College, Amherst, MA 01002 , U .S.A. Received September 6,1 991 ; accepted March 31 ,1992.
Aging Clin. Exp. Res., Vol. 4 , N. 3 219
A.P. Woodhull-McNeal
body falls very close to the midpoint of the base of support, so that stability is maintained (13). The balance about the knee and hip joints in young adults also maximizes stability. Viewed in the sagittal plane, the centers of gravity of the portions of the body above the knee and above the hip fall nearly upon these joints, but on average a little in front of the knee and behind the hip, hence tending to hyperextend each joint and increase its stability (12 ,14) ; this balanced alignment means that postural muscles need only to exert minimal force in order to maintain upright posture. No data have been published concerning agerelated changes in postural parameters such as location of joint centers and the body center of gravity. Therefore, this study examined the alignment of joint centers and measures of whole-body and segmental balance in 41 female subjects, aged 65 and older. It was hypothesized that older subjects would be less well-aligned and show greater segmental imbalances, compared to a younger reference population . Because the whole-body center of gravity is tightly controlled, it was predicted that the center of gravity would not change with age.
MATERIALS AND METHODS .Subjects were recruited at a senior center in Amherst, and at a public housing complex for the. elderly in Northampton, Massachusetts. All subjects were Caucasian women living independently;' most of them in apartments for the elderly. All were ambulatory, and none had lost limbs, although three had hip replacements. Most of them were quite mobile, but an effort was made also to recruit less active persons, in part by testing at the place of residence. No subject had Parkinson's diselise. Each subject was paid a small honorarium and gave written informed consent. Subjects filled out a questionnaire, with help if necessary. The items included age, previous maximum height , questions about back pain , osteoporosis, number of falls in the previous year, and activity. .Items concerning previous maximum height and osteoporosis are open to possible error in reporting. The question on back pain simply asked about the degree of back pain , with no attempt to ascribe a particular cause.
220 Aging Clin. Exp. Res.. Vol. 4, N. 3
Activity was self-rated in two ways: a relative activity scale and a more objective physical activity scale. Relative activity was a five-point subjective rating of "activity in relation to others your age" on a scale from "much less active" through "about the same as others" to "much more active." "About the same as others" was assigned a value of zero . Physical activity was assessed through a series of questions about activities such as chores around the home , yard work , blocks walked per week, and other activities. The responses on the scale questions were combined into a measure in which , for example, a subject who walked 5 to 10 blocks per week or walked 2 to 5 blocks and also exercised received a score of five, and one who walked 2 miles or more per week received a score of seven. The maximum possible score was eight , minimum zero. Although this scale was not validated, it is similar to others used for the elderly (15) and the general population (16), and gives a rough idea of the subjects ' level of physical activity. Subjects' heights and weights were measured. Markers for the joint centers of knee and hip were attached by elastic straps around the joints; the locations of all markers were determined by palpation by the same investigator. Subjects were photographed in profile on a simple, portable unidirectional force platform (reaction board) with the display visible in the photograph (12) (Fig . lA). A measure of kyphosis was taken as the angle on the photograph from the back of the waist to the most posterior point of the kyphosis to the last cervical vertebra. Data on joint center locations were digitized from the photographs. The center of gravity was calculated from the ratio of the scale reading to the subject's total weight, and was located relative to the ankle from the digitized data . This method strictly measures the center of pressure, but this differs from the center of gravity by only millimeters during quiet standing (17). The partial centers of gravity above the knee and hip were calculated by the method of Woodhull et al. (12), in which the quantity (mass X center of gravity) for each lower body part is subtracted from the whole-body center of gravity. The resulting "partial center of graVity" above the hip, for example, indicates how the body's weight is balanced above the hip , with positive numbers
Chan ges in posture and balance
corresponding to a forward-leaning moment. Data analysis was done with SPSSX , using analysis of variance (ANOVA), two-tailed T-tests , Ftests for difference in variance, and Pearson correlations as indicated in the text ; the p=0.05 level of significance was used. Comparisons were made with a previously-studied college-age population (12).
RESULTS Basic information on subjects is given in Table 1. On the whole , this was a fairly active group, with the physical activity scale mean of 5.7 units indicating more than 5 to 10 blocks walked per week. About half the subjects of each age group reported falling at least once in the past year. Of those who could recall their earlier maximum height, these subjects had lost on average 5 em due to aging , although only 27 % of them (and 32% of all subjects) reported having osteoporosis. In order to examine trends within this group , the elderly subjects were divided into "old" (65-79 years) and "very old" (80 or more) subgroups (Table 1). Compared to the "old" subjects, the "very old" subjects were significantly lighter in weight and fewer reported having osteoporosis, although their average change in height was just as large . Older subjects rated themselves as
more active relative to their peers , while their activity scores on the physical activity scale were not significantly different from the 65-79 age group. In the postural measures analyzed below, there were no significant differences between "old" and "very old" subjects in this study, except that the "very old" carried their heads further forward (ears 9 .2 em forward compared to 5.8 cm in the "old" group, t-test p = 0.003). The mean posture of these older subjects differed significantly from a college-age comparison group in three respects: the hips were on average further back; the average center of gravity above the hips was further forward (greater forward lean); and the kyphosis was greater (kyphosis angle smaller) than in the younger subjects (Table 2). The overall center of gravity was not changed at all. That is, the older subjects on average tended to lean forward from the hips, while maintaining the same overall anterior-posterior balance. Figure 1 shows profiles of the typical younger and older subjects. Two clear differences are in the greater kyphosis of the older subject and the greater forward lean from the hips with compensation in the hip position, which is more posterior than in the younger subject. However, the older population showed great variability, larger than among the young (Table 2). Manyolder subjects showed upright posture more similar
Table 1 - Basic data on e lde rly population studied. Means for each age range are given (SD in parentheses). "Charige in bt " is previous max imum reported height minus current measured height (numbers are given in brackets since data are incomplete). "Kyp h " is kyphosis angl e; smaller angles are great er kyphosis. "Re i act iv" is sub ject ive self-rat ed actiVity relative to peers of same age. "Acti vity " is scored f rom reported physical actiVity per week. "Back pain" 'is percent reportin g moderate to severe back pain . "Falls" is percent reporting any falls in past year. "Osteop " is percent reporting having osteoporosis. Age range
N
Ht (cm)
Wt (kg)
Change in ht (cm)
Kyph (deg)
ReI Activ
Activity
Back pain
Falls
Osteop
65-79
26
158 (6)
67 (10)
4 .8 (3.5) 1171
141 (7)
0.2 (1.1)
5.5 (1.6)
42 %
52%
46 %
80+
15
155 (7)
58 (8)
5.4 (2.9)
137 (7)
1.2 (0.9)
5.4 (1.8)
27%
47%
8%
[11]
• indicates significant differences between age groups (p < 0 .0 5 by t-test with poo led variances for weight and relative activity; by X2 for osteo porosis).
Aging Clin. Exp, Res ., Vol. 4 , N. 3 221
A.P. Wo odhull·McNeal
Table 2 - Postural data for 41 old er and 15 college-age subjects . All measures are given in cm . The positions of kn ee, hip, shoulder and ear are cm anterior to the ankle joint. Th e cente rs of gravity of the portions of the bod y abo ve the kn ee and hip are calculated as given in Method s. p values fo r mean d ifferences are f rom t·test, pool ed variances; pvalues fo r difference in variances are f rom F-test . Older subjects Mean (SO)
College-age subjects Mean (SO)
Knee (ern)
2.8 (2.1)
Hip
p-value for mean difference
p-value for difference in variances
3.8 (2.0)
NS
NS
4.6 (2.4)
6.1 (1.3)
0.03
0.015
Sh oulder
2.5 (3.5)
3.9 (1.9)
NS
0.017
Ear
7.0 (3.7)
5.9 (1.7)
NS
0.002
Whole body
4.5 (1.6)
4.9 (1.3)
NS
NS
Above knee
2.0 (2.3)
1.4 (1.1)
NS
0.004
Above hip
0.6 (2.5)
-1.0 (1.5)
0.0 3
0.057
139 (7.3)
143 (5.3)
0.03
NS
Position of
Center of gravity
Kyphosis
to that of younger people (Fig. 2A). Two other patterns were: large kyphosis but no forward leaning (Fig. 2B); and large kyphosis with great forward lean (Fig. 2C), which tended to be seen among subjects who appeared subjectively more unsteady. Statistical analysis bore out the impression that postural measures varied independently of one another; there was no significant correlation between kyphosis and forward leaning from the hips (R = 0 .096, p = 0 .280), nor between forward leaning and whole-body center of gravity (R = 0 .155, p = 0 .173), nor between kyphosis and center of gravity (R = 0 .015, p = 0.464). Subjects who were more active tended to have less forward lean. The center of gravity above the hip correlated negatively with the activity scale (R = - 0 .280, p = 0.04) and nearly reached significance correlated with self-rated activity relative to peers (R = - 0 .24, p = 0 .06). Attempts were made to relate history of falling
222 Aging Clin. Exp . Res., Vol. 4 , N. 3
to each of the postural variables. Discriminant analysis showed that none of the variables (center of gravity, kyphosis, forward lean, position of hips, shoulders, and knees) was significantly related to fall history. DISCUSSION
Three main differences are seen in elderly subjects compared to younger ones: increased kyphosis , decreased height , and a forward lean from the hip (more anterior center of gravity above the hips). On the average , these changes were all observed in our study population, relative to the younger group (Table 2 and Fig. 1). However, these changes do not necessarily appear consistently together, and the range of postures among the elderly, even in this mobile population , is quite large (Table 2 and Fig. 2). Brocklehurst et al. (10) also found that there was not one consistent complex of postural
Changes in posture and balance
A
B
PLUMB LINE
I
SCALE
I
Figure 1 - Typical subjects from the young and old age groups, whose joint centers and centers of gravity are very close to the group means (Table 2). A. Collegeage. also showing measurement system and placement of markers. The vertical line extending from the ankle is for reference. B. Over-65 typical subject.
A
B
changes in the elderly. They found that kyphosis was positively correlated with change in height, negatively correlated with weight, and negatively correlated to grip strength. The first of these findings is corroborated by this study. These workers also found that flexion at the knee was increased, but did not correlate to change in height, weight, or grip strength; instead, it was related to increased sway. Similarly, this study finds increased leaning unrelated to other changes, but related to activity. Increased kyphosis and decreased height are seen together, probably due to osteoporosis affecting the thoracic vertebrae (1). On the other hand, if compression fractures occur further down in the lumbar vertebrae, the decrease in height may not be as dramatic, and there may be very little change in observable posture. Fig. 2A illustrates a subject who reported having osteoporosis with vertebral fractures, but whose posture is very youthful. The finding that many more of the group aged 65-79 reported osteoporosis, compared to those over 80, probably reflects differences in awareness and relative exposure to information on osteoporosis, rather than a real difference
c
Figure 2 - Several patterns of posture found in the elderly group. A. "youthful" posture despite reported vertebral fractures. B. Kyphosis but no forward lean from the hips. C. Kyphosis and forward lean.
Aging Clin. Exp. Res., Vol. 4, N. 3 223
A.P. Woodhull-McNeal
in incidence. No objective diagnosis was made, and the change in height for the older group was as great as that for the younger group. The tendency to lean forward from the hips is statistically related to activity level, with more active seniors being more upright. This perhaps suggests that those who have greater muscular strength lean less. Alternatively, those whose frailty or instability lead them to lean forward for security may also tend to avoid major activity. Leaning forward with the upper body involves compensation of other body parts in order to keep the overall center of gravity constant (18). Thus, in elders who lean forward, the hips are located further back so that the weight remains balanced above the feet. The overall center of gravity is remarkably stable, being maintained midway along the base of support despite these postural changes; likewise, the center of gravity is maintained when weight is carried in a backpack (19). However, these compensations may have their cost in terms of muscle use, since the forward-leaning posture is far from mechanical equilibrium, so muscle groups in the back and around the hip must be continually more active to keep the trunk from further forward rotation. Because the subjects were volunteers, they may represent the more socially and physically active segment of the over-65 population in this area. Their self-ratings of activity relative to peers, especially for the over-80 group, would tend to corroborate this. On the other hand, the incidence of falling (50% overall) was similar to that reported in the more extensive community survey by Blake et al., who found that 42% of the women surveyed had fallen in the preceding year (20). The major finding of this paper is the great variability of posture among the elderly, reflecting the many ways in which the musculo-skeletal system can age depending on activity, osteoporosis, and other health conditions. There is a suggestion that different processes lead to kyphosis than lead to forward leaning, since kyphosis is related to decreased height, whereas forward lean correlates with lower levels of activity. Further investigation into the question of forward leaning could address whether it is caused by changes in lumbar vertebrae or is an adaptive response to postural instability or other factors.
224 Aging Clin. Exp. Res., Vol. 4, N. 3
ACKNO~DGEMENTS
This work was supported in part by faculty development grants to Hampshire College from the Andrew W. Mellon Foundation, and the William and Flora Hewlett Foundation. I am very grateful to all the volunteers who made this study possible. I also want to thank Esther Clark for her substantial aid in recruiting subjects, and Louise Gillette for her cheerful and skillful help in testing subjects. Gordon Woodhull gave able technicalassistancewith the digitizing program.
REFERENCES 1. Giansiracusa D.F., Kantrowitz F.G.: Rheumatic and Metabolic Bone Diseases in the Elderly. Collamore Press, Lexington MA, 1982. 2. Woollacott M.H., Shumway-Cook A, Nashner L.M.: Aging and posture control: changes in sensory organization and muscular coordination. Int. J. Aging Hum. Dev. 23: 97-114,1986. 3. Hayes KC., Spencer J.D., Riach c.i., Lucy S.D., Kirshen AJ.: Age-related changes in postural sway. In: Winter D.A, et al. (Eds.), Biomechanics IX-A. Human Kinetics Press, Champaign, IL, 1985, pp. 383-387. 4. Norre M.E., Forrez G., Beckers A: Posturography measuring instability in vestibular dysfunctioning in the elderly. Age Ageing 16: 89-93,1987. 5. Richardson L.D., Roth J.H., Vandervoort AA, Hayes KC., Mac Kenzie R.A.: Postural stability and Colles' fracture. Age Ageing 16: 133-138, 1987. 6. Era P., Heikkinen E.: Postural sway during standing and unexpected disturbance of balance in random samples of men of different ages. J. Gerontol. 40: 287295,1985. 7. Fernie GR, Gryfe cr. Holliday P.J., Llewellyn A: The relationship of postural sway in standing to the incidence of falls in geriatric subjects. Age Ageing 11: 11-16,1982. 8. Imms F.J., Edholm a.G.: Studies of gait and mobility in the elderly. Age Ageing 10: 147-156, 1981. 9. Wolfson L., Whipple R., Amerman P., Tobin J.N.: Gait assessment in the elderly: a gait abnormality rating scale and its relation to falls. J. Gerontol. 45: M12-19, 1990. 10. Brocklehurst .i.c., Robertson D., James-Groom P.: Skeletal deformities in the elderly and their effect on postural sway. J. Am. Geriatr. Soc. 30: 534-538, 1982. 11. Wells KF., Luttgens K: Kinesiology: Scientific Basis of Motion. Saunders, Philadelphia PA, 1976. 12. Woodhull AM., Maltrud K, Mello B.L.: Alignment of the human body in standing. Eur. J. Appl. Physiol. 54: 109-115,1985.
Changes in posture and balance
13 . Hellebrandt F.A.: Standing as a geotropic reflex: The mechanism of the asynchronous rotation of motor units. Am. J . Physiol. 121: 471-474, 1938. 14 . Akerblom B.: Standing and S itting Posture. A.-B. Nordiska Bokhandeln , Stockolm, 1948. 15. Voorrips L.E., Ravelli A.c.J. , Dongelmans P.C.A. , Deurenberg P., van Staveren W.A.: A physical activity questionnaire for the elderly. Med. Sci. Sports Exere. 23 : 974-979,1991. 16. Paffenbarger R.S. , Wing A.L., Hyde R.T.: Physical activity as an index of heart attack risk in college alumni. Am. J. Epidemiol. 108: 161-165, 1978.
17 . Winter D.A., Patla A.E., Frank J .S.: Assessment of balance control in humans . Med. Prog. Techno/. 16 : 31 -51 ,1990. 18. Hay J .G.: The center of gravity of the human body. Kines iology 3: 20-44, 1973. 19 . Bloom D., Woodhull-McNeal Ann P.: Postural adjustments while standing with two types of loaded backpack. Ergonomics 30 : 1425-1430, 1987. 20 . Blake A.J ., Morgan K., Bendall M.J. , Dallosso H., Ebrahim S.B.J., Arie T.H .D., Fentem P.H ., Bassey E.J.: Falls by elderly people at home: prevalence and associated factors. Age Ageing 17 : 365-372, 1988.
Aging Clin. Exp. Res., Vol. 4, N. 3 225
Changes in posture and balance with age A.P. Woodhull-McNeal School of Natural Science, Hampshire College, Amherst, Massachusetts, U.S.A. ABSTRACT. Changes in posture with age are of concern because of their association with impaired mobility and the possibility of falls. In this study alignment of the joint centers and balance of body segments were measured in 41 women , aged 65 and older, and correlated with weight, age, change in height, and activity. Balance was assessed by calculating the whole-body center of gravity, and the partial centers of gravity above the knee and hip. Compared to a younger reference population, the older subjects had greater kyphosis , a more posterior hip position , and leaned forward more from the hips (more anterior center of gravity above the hips). Among the 41 elderly subjects, larger kyphoses correlated with greater decrease in height. Elderly subjects who were inactive tended to lean more, but forward lean was not correlated to other variables. These findings suggest two separate sets of changes: an osteoporosis-related stooping, and also a forward lean that may relate to muscular weakness or to fear of instability. Posture among the elderly was quite varied and did not follow anyone pattern of change. (Aging Clin . Exp. Res. 4: 219-225, 1992) INTRODUCTION Among the numerous changes associated with aging, changes in posture seem to signal major changes in function. The stereotypical picture of old age is one of a stooped-over torso with exaggerated kyphosis, forward-leaning upper body, and bent knees . Osteoporosis can cause some of these changes, since compression fractures lead
to "wedging" of the anterior side of the vertebrae, causing both increased kyphosis and decreased lumbar lordosis (1); the lumbar changes may in turn lead to changes in pelvic tilt and knee position . Arthritis, muscle weakness, and other conditions may also contribute to different postural changes. Most studies concerning posture in the elderly have focussed on postural sway. (2-5). This is a logical concern since increased sway may be a sign of generalized instability, and thus a sign of increased danger of falling. Not surprisingly , sway has been found to increase with age (4, 6), and old people who fall have on average increased postural sway (5, 7). Changes in gait are another major focus of research (8). It was suggested that increased forward leaning of head , arms and trunk (HAT) is part of a complex of changes in gait associated with instability (9). Standing posture, including the static balance of body segments, has been studied very little, although it could be important in understanding some of the reasons for postural instability. When posture changes , then the biomechanics of standing and of locomotion must also change, which could add to mechanical instability and insecurity. Brocklehurst et al. (10) correlated some measures of skeletal change to postural sway, and found that increased sway was associated with larger loss of height, greater knee flexion, and less grip strength. In young adults the skeleton is aligned for maximal stability in several ways. Although not strictly linear, the joint centers are aligned fairly linearly above and forward of the ankle joint (11 , 12) . The center of gravity of the whole
Key words : Aging, balance, posture. Correspondence: Ann P. Woodhull-McNeal, PhD., School of Natural Science, Hampshire College, Amherst, MA 01002 , U .S.A. Received September 6,1 991 ; accepted March 31 ,1992.
Aging Clin. Exp. Res., Vol. 4 , N. 3 219
A.P. Woodhull-McNeal
body falls very close to the midpoint of the base of support, so that stability is maintained (13). The balance about the knee and hip joints in young adults also maximizes stability. Viewed in the sagittal plane, the centers of gravity of the portions of the body above the knee and above the hip fall nearly upon these joints, but on average a little in front of the knee and behind the hip, hence tending to hyperextend each joint and increase its stability (12 ,14) ; this balanced alignment means that postural muscles need only to exert minimal force in order to maintain upright posture. No data have been published concerning agerelated changes in postural parameters such as location of joint centers and the body center of gravity. Therefore, this study examined the alignment of joint centers and measures of whole-body and segmental balance in 41 female subjects, aged 65 and older. It was hypothesized that older subjects would be less well-aligned and show greater segmental imbalances, compared to a younger reference population . Because the whole-body center of gravity is tightly controlled, it was predicted that the center of gravity would not change with age.
MATERIALS AND METHODS .Subjects were recruited at a senior center in Amherst, and at a public housing complex for the. elderly in Northampton, Massachusetts. All subjects were Caucasian women living independently;' most of them in apartments for the elderly. All were ambulatory, and none had lost limbs, although three had hip replacements. Most of them were quite mobile, but an effort was made also to recruit less active persons, in part by testing at the place of residence. No subject had Parkinson's diselise. Each subject was paid a small honorarium and gave written informed consent. Subjects filled out a questionnaire, with help if necessary. The items included age, previous maximum height , questions about back pain , osteoporosis, number of falls in the previous year, and activity. .Items concerning previous maximum height and osteoporosis are open to possible error in reporting. The question on back pain simply asked about the degree of back pain , with no attempt to ascribe a particular cause.
220 Aging Clin. Exp. Res.. Vol. 4, N. 3
Activity was self-rated in two ways: a relative activity scale and a more objective physical activity scale. Relative activity was a five-point subjective rating of "activity in relation to others your age" on a scale from "much less active" through "about the same as others" to "much more active." "About the same as others" was assigned a value of zero . Physical activity was assessed through a series of questions about activities such as chores around the home , yard work , blocks walked per week, and other activities. The responses on the scale questions were combined into a measure in which , for example, a subject who walked 5 to 10 blocks per week or walked 2 to 5 blocks and also exercised received a score of five, and one who walked 2 miles or more per week received a score of seven. The maximum possible score was eight , minimum zero. Although this scale was not validated, it is similar to others used for the elderly (15) and the general population (16), and gives a rough idea of the subjects ' level of physical activity. Subjects' heights and weights were measured. Markers for the joint centers of knee and hip were attached by elastic straps around the joints; the locations of all markers were determined by palpation by the same investigator. Subjects were photographed in profile on a simple, portable unidirectional force platform (reaction board) with the display visible in the photograph (12) (Fig . lA). A measure of kyphosis was taken as the angle on the photograph from the back of the waist to the most posterior point of the kyphosis to the last cervical vertebra. Data on joint center locations were digitized from the photographs. The center of gravity was calculated from the ratio of the scale reading to the subject's total weight, and was located relative to the ankle from the digitized data . This method strictly measures the center of pressure, but this differs from the center of gravity by only millimeters during quiet standing (17). The partial centers of gravity above the knee and hip were calculated by the method of Woodhull et al. (12), in which the quantity (mass X center of gravity) for each lower body part is subtracted from the whole-body center of gravity. The resulting "partial center of graVity" above the hip, for example, indicates how the body's weight is balanced above the hip , with positive numbers
Chan ges in posture and balance
corresponding to a forward-leaning moment. Data analysis was done with SPSSX , using analysis of variance (ANOVA), two-tailed T-tests , Ftests for difference in variance, and Pearson correlations as indicated in the text ; the p=0.05 level of significance was used. Comparisons were made with a previously-studied college-age population (12).
RESULTS Basic information on subjects is given in Table 1. On the whole , this was a fairly active group, with the physical activity scale mean of 5.7 units indicating more than 5 to 10 blocks walked per week. About half the subjects of each age group reported falling at least once in the past year. Of those who could recall their earlier maximum height, these subjects had lost on average 5 em due to aging , although only 27 % of them (and 32% of all subjects) reported having osteoporosis. In order to examine trends within this group , the elderly subjects were divided into "old" (65-79 years) and "very old" (80 or more) subgroups (Table 1). Compared to the "old" subjects, the "very old" subjects were significantly lighter in weight and fewer reported having osteoporosis, although their average change in height was just as large . Older subjects rated themselves as
more active relative to their peers , while their activity scores on the physical activity scale were not significantly different from the 65-79 age group. In the postural measures analyzed below, there were no significant differences between "old" and "very old" subjects in this study, except that the "very old" carried their heads further forward (ears 9 .2 em forward compared to 5.8 cm in the "old" group, t-test p = 0.003). The mean posture of these older subjects differed significantly from a college-age comparison group in three respects: the hips were on average further back; the average center of gravity above the hips was further forward (greater forward lean); and the kyphosis was greater (kyphosis angle smaller) than in the younger subjects (Table 2). The overall center of gravity was not changed at all. That is, the older subjects on average tended to lean forward from the hips, while maintaining the same overall anterior-posterior balance. Figure 1 shows profiles of the typical younger and older subjects. Two clear differences are in the greater kyphosis of the older subject and the greater forward lean from the hips with compensation in the hip position, which is more posterior than in the younger subject. However, the older population showed great variability, larger than among the young (Table 2). Manyolder subjects showed upright posture more similar
Table 1 - Basic data on e lde rly population studied. Means for each age range are given (SD in parentheses). "Charige in bt " is previous max imum reported height minus current measured height (numbers are given in brackets since data are incomplete). "Kyp h " is kyphosis angl e; smaller angles are great er kyphosis. "Re i act iv" is sub ject ive self-rat ed actiVity relative to peers of same age. "Acti vity " is scored f rom reported physical actiVity per week. "Back pain" 'is percent reportin g moderate to severe back pain . "Falls" is percent reporting any falls in past year. "Osteop " is percent reporting having osteoporosis. Age range
N
Ht (cm)
Wt (kg)
Change in ht (cm)
Kyph (deg)
ReI Activ
Activity
Back pain
Falls
Osteop
65-79
26
158 (6)
67 (10)
4 .8 (3.5) 1171
141 (7)
0.2 (1.1)
5.5 (1.6)
42 %
52%
46 %
80+
15
155 (7)
58 (8)
5.4 (2.9)
137 (7)
1.2 (0.9)
5.4 (1.8)
27%
47%
8%
[11]
• indicates significant differences between age groups (p < 0 .0 5 by t-test with poo led variances for weight and relative activity; by X2 for osteo porosis).
Aging Clin. Exp, Res ., Vol. 4 , N. 3 221
A.P. Wo odhull·McNeal
Table 2 - Postural data for 41 old er and 15 college-age subjects . All measures are given in cm . The positions of kn ee, hip, shoulder and ear are cm anterior to the ankle joint. Th e cente rs of gravity of the portions of the bod y abo ve the kn ee and hip are calculated as given in Method s. p values fo r mean d ifferences are f rom t·test, pool ed variances; pvalues fo r difference in variances are f rom F-test . Older subjects Mean (SO)
College-age subjects Mean (SO)
Knee (ern)
2.8 (2.1)
Hip
p-value for mean difference
p-value for difference in variances
3.8 (2.0)
NS
NS
4.6 (2.4)
6.1 (1.3)
0.03
0.015
Sh oulder
2.5 (3.5)
3.9 (1.9)
NS
0.017
Ear
7.0 (3.7)
5.9 (1.7)
NS
0.002
Whole body
4.5 (1.6)
4.9 (1.3)
NS
NS
Above knee
2.0 (2.3)
1.4 (1.1)
NS
0.004
Above hip
0.6 (2.5)
-1.0 (1.5)
0.0 3
0.057
139 (7.3)
143 (5.3)
0.03
NS
Position of
Center of gravity
Kyphosis
to that of younger people (Fig. 2A). Two other patterns were: large kyphosis but no forward leaning (Fig. 2B); and large kyphosis with great forward lean (Fig. 2C), which tended to be seen among subjects who appeared subjectively more unsteady. Statistical analysis bore out the impression that postural measures varied independently of one another; there was no significant correlation between kyphosis and forward leaning from the hips (R = 0 .096, p = 0 .280), nor between forward leaning and whole-body center of gravity (R = 0 .155, p = 0 .173), nor between kyphosis and center of gravity (R = 0 .015, p = 0.464). Subjects who were more active tended to have less forward lean. The center of gravity above the hip correlated negatively with the activity scale (R = - 0 .280, p = 0.04) and nearly reached significance correlated with self-rated activity relative to peers (R = - 0 .24, p = 0 .06). Attempts were made to relate history of falling
222 Aging Clin. Exp . Res., Vol. 4 , N. 3
to each of the postural variables. Discriminant analysis showed that none of the variables (center of gravity, kyphosis, forward lean, position of hips, shoulders, and knees) was significantly related to fall history. DISCUSSION
Three main differences are seen in elderly subjects compared to younger ones: increased kyphosis , decreased height , and a forward lean from the hip (more anterior center of gravity above the hips). On the average , these changes were all observed in our study population, relative to the younger group (Table 2 and Fig. 1). However, these changes do not necessarily appear consistently together, and the range of postures among the elderly, even in this mobile population , is quite large (Table 2 and Fig. 2). Brocklehurst et al. (10) also found that there was not one consistent complex of postural
Changes in posture and balance
A
B
PLUMB LINE
I
SCALE
I
Figure 1 - Typical subjects from the young and old age groups, whose joint centers and centers of gravity are very close to the group means (Table 2). A. Collegeage. also showing measurement system and placement of markers. The vertical line extending from the ankle is for reference. B. Over-65 typical subject.
A
B
changes in the elderly. They found that kyphosis was positively correlated with change in height, negatively correlated with weight, and negatively correlated to grip strength. The first of these findings is corroborated by this study. These workers also found that flexion at the knee was increased, but did not correlate to change in height, weight, or grip strength; instead, it was related to increased sway. Similarly, this study finds increased leaning unrelated to other changes, but related to activity. Increased kyphosis and decreased height are seen together, probably due to osteoporosis affecting the thoracic vertebrae (1). On the other hand, if compression fractures occur further down in the lumbar vertebrae, the decrease in height may not be as dramatic, and there may be very little change in observable posture. Fig. 2A illustrates a subject who reported having osteoporosis with vertebral fractures, but whose posture is very youthful. The finding that many more of the group aged 65-79 reported osteoporosis, compared to those over 80, probably reflects differences in awareness and relative exposure to information on osteoporosis, rather than a real difference
c
Figure 2 - Several patterns of posture found in the elderly group. A. "youthful" posture despite reported vertebral fractures. B. Kyphosis but no forward lean from the hips. C. Kyphosis and forward lean.
Aging Clin. Exp. Res., Vol. 4, N. 3 223
A.P. Woodhull-McNeal
in incidence. No objective diagnosis was made, and the change in height for the older group was as great as that for the younger group. The tendency to lean forward from the hips is statistically related to activity level, with more active seniors being more upright. This perhaps suggests that those who have greater muscular strength lean less. Alternatively, those whose frailty or instability lead them to lean forward for security may also tend to avoid major activity. Leaning forward with the upper body involves compensation of other body parts in order to keep the overall center of gravity constant (18). Thus, in elders who lean forward, the hips are located further back so that the weight remains balanced above the feet. The overall center of gravity is remarkably stable, being maintained midway along the base of support despite these postural changes; likewise, the center of gravity is maintained when weight is carried in a backpack (19). However, these compensations may have their cost in terms of muscle use, since the forward-leaning posture is far from mechanical equilibrium, so muscle groups in the back and around the hip must be continually more active to keep the trunk from further forward rotation. Because the subjects were volunteers, they may represent the more socially and physically active segment of the over-65 population in this area. Their self-ratings of activity relative to peers, especially for the over-80 group, would tend to corroborate this. On the other hand, the incidence of falling (50% overall) was similar to that reported in the more extensive community survey by Blake et al., who found that 42% of the women surveyed had fallen in the preceding year (20). The major finding of this paper is the great variability of posture among the elderly, reflecting the many ways in which the musculo-skeletal system can age depending on activity, osteoporosis, and other health conditions. There is a suggestion that different processes lead to kyphosis than lead to forward leaning, since kyphosis is related to decreased height, whereas forward lean correlates with lower levels of activity. Further investigation into the question of forward leaning could address whether it is caused by changes in lumbar vertebrae or is an adaptive response to postural instability or other factors.
224 Aging Clin. Exp. Res., Vol. 4, N. 3
ACKNO~DGEMENTS
This work was supported in part by faculty development grants to Hampshire College from the Andrew W. Mellon Foundation, and the William and Flora Hewlett Foundation. I am very grateful to all the volunteers who made this study possible. I also want to thank Esther Clark for her substantial aid in recruiting subjects, and Louise Gillette for her cheerful and skillful help in testing subjects. Gordon Woodhull gave able technicalassistancewith the digitizing program.
REFERENCES 1. Giansiracusa D.F., Kantrowitz F.G.: Rheumatic and Metabolic Bone Diseases in the Elderly. Collamore Press, Lexington MA, 1982. 2. Woollacott M.H., Shumway-Cook A, Nashner L.M.: Aging and posture control: changes in sensory organization and muscular coordination. Int. J. Aging Hum. Dev. 23: 97-114,1986. 3. Hayes KC., Spencer J.D., Riach c.i., Lucy S.D., Kirshen AJ.: Age-related changes in postural sway. In: Winter D.A, et al. (Eds.), Biomechanics IX-A. Human Kinetics Press, Champaign, IL, 1985, pp. 383-387. 4. Norre M.E., Forrez G., Beckers A: Posturography measuring instability in vestibular dysfunctioning in the elderly. Age Ageing 16: 89-93,1987. 5. Richardson L.D., Roth J.H., Vandervoort AA, Hayes KC., Mac Kenzie R.A.: Postural stability and Colles' fracture. Age Ageing 16: 133-138, 1987. 6. Era P., Heikkinen E.: Postural sway during standing and unexpected disturbance of balance in random samples of men of different ages. J. Gerontol. 40: 287295,1985. 7. Fernie GR, Gryfe cr. Holliday P.J., Llewellyn A: The relationship of postural sway in standing to the incidence of falls in geriatric subjects. Age Ageing 11: 11-16,1982. 8. Imms F.J., Edholm a.G.: Studies of gait and mobility in the elderly. Age Ageing 10: 147-156, 1981. 9. Wolfson L., Whipple R., Amerman P., Tobin J.N.: Gait assessment in the elderly: a gait abnormality rating scale and its relation to falls. J. Gerontol. 45: M12-19, 1990. 10. Brocklehurst .i.c., Robertson D., James-Groom P.: Skeletal deformities in the elderly and their effect on postural sway. J. Am. Geriatr. Soc. 30: 534-538, 1982. 11. Wells KF., Luttgens K: Kinesiology: Scientific Basis of Motion. Saunders, Philadelphia PA, 1976. 12. Woodhull AM., Maltrud K, Mello B.L.: Alignment of the human body in standing. Eur. J. Appl. Physiol. 54: 109-115,1985.
Changes in posture and balance
13 . Hellebrandt F.A.: Standing as a geotropic reflex: The mechanism of the asynchronous rotation of motor units. Am. J . Physiol. 121: 471-474, 1938. 14 . Akerblom B.: Standing and S itting Posture. A.-B. Nordiska Bokhandeln , Stockolm, 1948. 15. Voorrips L.E., Ravelli A.c.J. , Dongelmans P.C.A. , Deurenberg P., van Staveren W.A.: A physical activity questionnaire for the elderly. Med. Sci. Sports Exere. 23 : 974-979,1991. 16. Paffenbarger R.S. , Wing A.L., Hyde R.T.: Physical activity as an index of heart attack risk in college alumni. Am. J. Epidemiol. 108: 161-165, 1978.
17 . Winter D.A., Patla A.E., Frank J .S.: Assessment of balance control in humans . Med. Prog. Techno/. 16 : 31 -51 ,1990. 18. Hay J .G.: The center of gravity of the human body. Kines iology 3: 20-44, 1973. 19 . Bloom D., Woodhull-McNeal Ann P.: Postural adjustments while standing with two types of loaded backpack. Ergonomics 30 : 1425-1430, 1987. 20 . Blake A.J ., Morgan K., Bendall M.J. , Dallosso H., Ebrahim S.B.J., Arie T.H .D., Fentem P.H ., Bassey E.J.: Falls by elderly people at home: prevalence and associated factors. Age Ageing 17 : 365-372, 1988.
Aging Clin. Exp. Res., Vol. 4, N. 3 225
