METHODOLOGY
Subjects
Subjects were Filipino men and women 60 years old and above who have participated in the 5th National Nutrition Survey in 1998. There were 4,541 adults, out of whom 33.6% were 60 years old and above. The study covered 97 geographic areas which included 77 provinces in the 15 regions, 5 cities, 5 center areas in the National Capital Region (NCR) and 10 highly urbanized cities (HUC). A two-stage stratified sampling design was employed, with the barangay as the primary unit and the individual as the secondary unit.
Anthropometric Measurements
The knee height was measured using a sliding broad-blade caliper (Ross). While lying supine, the subjects were asked to bend both the left knee and left ankle at 90o. The fixed blade was placed under the heel of the left foot while the sliding blade was pressed down against the thigh about 5.1 cm proximal to the knee cap (patella). The shaft of the caliper was in line with long bone in the lower leg (tibia) and was over the ankle bone (lateral malleolus). The locking lever was pushed away from the blades to hold the measurement and was read through the viewing window to the nearest 0.1 cm. Two measurements were made in immediate succession which should agree within 0.5 cm (3).
Data and Analysis
Descriptive statistics were used in sorting anthropometric characteristics of subjects into four age groups and by gender. T-tests were used to determine the significance of differences in measurements of males and females. Multiple regression was utilized to examine the relationship of stature to knee height, arm span and age. Pearson product moment correlation coefficients (r) between stature and knee height, arm span, and age were computed. All statistical analyses were performed using the Stata version 6.
RESULTS
A total of 1,527 Filipino adults 60 years old and above were included in the study, out of which, 51% were males and 49% were females. The subjects were divided into four age groups. Forty six (46%) of the study sample belonged to the 60-65 year age group, 27% to the 66-70 year age group and the remaining 27% comprised the > 70 and above age groups (Table 1). In general, male subjects were taller than female subjects using both standing height and the proxy indicators, knee height and arm span.
Table 2. Mean height, knee height, and arm span of the Filipino elderly of the FNRI 5th NNS
Age Group | Height(cm) | Knee Height (cm) | Arm Span (cm) |
Male | Female | Male | Female | Male | Female |
x | SD | x | SD | x | SD | x | SD | x | SD | x | SD |
| 60-65 | 160.5 | 6.3 | 149.6 | 5.2 | 49.4 | 2.6 | 46.0 | 2.4 | 166.4 | 7.3 | 154.1 | 6.4 |
| 66-70 | 158.1 | 5.6 | 147.1 | 5.5 | 48.5 | 2.4 | 45.4 | 2.5 | 164.1 | 6.5 | 152.2 | 7.0 |
| 71-75 | 159.3 | 5.3 | 147.9 | 5.5 | 49.6 | 2.9 | 45.7 | 2.3 | 159.1 | 5.5 | 154.0 | 7.1 |
| 76 & above | 157.8 | 5.7 | 145.4 | 6.0 | 48.9 | 2.7 | 45.1 | 2.4 | 165.2 | 7.3 | 152.0 | 7.0 |
| TOTAL | 159.3 | 6.0 | 148.1 | 5.6 | 49.1 | 2.7 | 45.7 | 2.4 | 164.4 | 9.3 | 153.3 | 6.8 |
Table 2 shows the mean height, knee height and arm span of the subjects. Male subjects had a mean height 159.3 cm while the females had a lower mean of 148.1 cm. Mean knee height and arm span were also higher among males (49.1 cm) than among females (45.7).
There was a 2.7 cm difference in the mean height of males aged 76 and older in comparison to their younger counterparts, 60-65 years of age, while the difference among females in the same age group was 4.2 cm.
Changes on knee height were not as pronounced as changes in height, showing only a 0.5 cm and 0.9 cm decrement for older males and females, respectively. For the age group 76 years and older, the mean knee heights of males and females were 48.9 cm and 45.0 cm, respectively and those of their younger counterparts (60-65 years) were 49.4 cm and 46.0 cm.
Closely paralleling the changes in mean height and knee height were the changes in arm span. The decrement of arm span of the older female subjects (2.1 cm) was close to twice than that of their male counterparts (1.2 cm). This pattern was also seen for height and knee height.
Regression equations for height estimation from knee height and arm span are presented in Table 3. In these equations, the knee height and arm span measurements are in centimeters and age is rounded off to the nearest whole year.
Table 4. Comparison of Actual Mean Height and Estimated Height from Knee Height
Age Group | Male | Female |
Actual Height | Estimated Height | P Value | Actual Height | Estimated Height | P Value |
x | SD | x | SD | | x | SD | x | SD | |
| 60-65 | 160.5 | 6.3 | 159.7 | 3.6 | 0.0412 | 149.6 | 5.2 | 149.4 | 3.7 | 0.5596 |
| 66-70 | 158.1 | 5.6 | 158.0 | 3.4 | 0.8230 | 147.1 | 5.5 | 147.6 | 3.8 | 0.3007 |
| 71-75 | 159.3 | 5.3 | 159.1 | 4.0 | 0.7568 | 147.9 | 5.5 | 147.2 | 3.6 | 0.2811 |
| 76 & above | 157.8 | 5.7 | 157.6 | 3.8 | 0.7763 | 145.4 | 6.0 | 145.2 | 3.7 | 0.7881 |
| TOTAL | 159.3 | 6.0 | 158.8 | 3.8 | 0.0525 | 148.1 | 5.6 | 148.1 | 4.0 | -1.0000 |
Table 4 shows the comparison between the actual mean height and estimated height from knee height. Estimated height values for both sexes showed no significant difference from the actual mean height of the subjects (male p<0.0525; female p< -1.0000).
Table 5. Comparison of Actual Mean Height and Estimated Height from Arm Span
Age Group | Male | Female |
Actual Height | Estimated Height | P Value | Actual Height | Estimated Height | P Value |
x | SD | x | SD | | x | SD | x | SD | |
| 60-65 | 160.5 | 6.3 | 160.4 | 2.0 | 0.7797 | 149.6 | 5.2 | 149.7 | 4.1 | 0.7786 |
| 66-70 | 158.1 | 5.6 | 159.4 | 1.8 | 0.0013 | 147.1 | 5.5 | 147.5 | 4.4 | 0.4318 |
| 71-75 | 159.3 | 5.3 | 157.7 | 4.3 | 0.0161 | 147.9 | 5.5 | 147.8 | 4.5 | 0.8877 |
| 76 & above | 157.8 | 5.7 | 158.9 | 2.1 | 0.0808 | 145.4 | 6.0 | 145.5 | 4.5 | 0.9000 |
| TOTAL | 159.3 | 6.0 | 159.5 | 2.7 | -0.8374 | 148.1 | 5.6 | 148.3 | 4.5 | 0.4524 |
Table 5 shows the comparison between the actual mean height of the subjects and the estimated height based on arm span. Values derived using the equation for both sexes showed no significant difference when compared with the actual mean height of the subjects (male p< -0.8374; female p<0.4524).
The use of the Japanese-American or Chumlea models to estimate height for Filipinos yielded overestimates as shown in Table 6. The Japanese-American equation significantly overestimated the stature of Filipino men and women respectively (male p<0.0000; female p<0.0000). The difference between the estimated height using both the Chumlea and Japanese-American equations and actual height of the subjects ranged from 1.4 cm to 5.0 cm among males and 2.5 cm to 4.2 cm among females.
Using the equation formulated in the present study, arm span and knee height were shown to be positively and strongly correlated with height in both male and female subjects, having a correlation value of 0.51 - 0.77 and 0.68 - 0.77, respectively (Table 7).
DISCUSSION
The use of knee height and arm span to estimate height of elderly subjects has been documented. Formulas were derived for both Caucasian and Japanese-American subjects for knee height. However, no local study using national data has been done. Thus, this study was undertaken to validate the use of knee height and arm span by developing an equation for height estimation for Filipino elderly (60 years and over).
The irregular patterns of decreasing trend in the mean height, knee height and arm span from ages 60-70 and an increase thereafter only indicate greater problems in the nutritional assessment of elderly persons. The presence of spinal curvature, which is more prevalent in the elderly than in younger persons may result in a spine which is so curved in advanced cases that the headboard cannot be positioned on the crown of the head for upright stature. In some elderly persons, the larger amount of adipose tissues on the buttocks prevent the heels, buttocks, shoulders and head from simultaneously touching the measuring scale.
Specifically for arm span, some problems that may be encountered are: (a) a person may not be able to stretch out both arms; (b) if the person's back is badly bent, one cannot measure arm span; and (c) if the person cannot straighten his/her finger because of a condition like arthritis.
While a greater decrease in stature in females than in males due to aging has been reported by Dequeker et. al. (9), this was not observed among the elderly black population (2) as well as in this study, indicating racial differences in the phenomenon.
It is noteworthy that knee height has been found to be lower in the older females than among younger ones except at 71-75 years, but no similar relationship was found in the male participants despite the lack of a physiologic basis for loss of knee height with age (10).
Chumlea, et. al. (4) found that knee height measured either in sitting or recumbent position, was highly correlated with stature in men and women and had a higher inter-observer reliability than arm span (4). In this study, although the decreasing trends in the knee height and arm span measurements were not clearly evident with advancing age, the regression equations and standard deviations of the proposed formula indicate a good relationship using the correlation coefficients between height and knee height
(r = 0.68 - 0.77) and between height and arm span (r = 0.51 - 0.77). The study of Verluis et. al. (11) in which height and arm span were found to decrease significantly with age showed the correlation between arm span and height to be 0.83.
Sullivan et. al. also showed that the reliability of knee height measurements is superior to that of arm length and other anthropometric variables (12). Nevertheless Mitchelle and Lipschitz suggested that arm length is a useful alternative to height measurements in elderly people (13).
In the serial study conducted by Chumlea et al of white, healthy, middle class, elderly men and women, aged 60-80 years, the rate of decrease in stature was estimated to be 0.5 cm/year. This rate of loss in stature is much greater than previously reported estimates and also approximately constant across cohorts (14).
Among the Japanese-American men, a quarter of the loss of height was explained by studies reporting lower bone mineral content and higher incidence of vertebral osteoporosis in this population (15).
The statistically significant differences in the predicted height of Filipino elderly using the equations developed in this study and the 2 racial models reviewed here may indicate different influences of aging in the three groups as well as differences in anatomical proportions in subjects over 60 years of age. Ethnic differences in length of body segments have also been documented. Chumlea et. al., working with the NHANES III data reported that non-Hispanic whites had significantly larger knee height and sitting height than did their Mexican American counterparts (2,14, 16, 17). Thus, Kohrs and Czajka, in their analysis of the NHANES data by race, reported some differences in stature and concluded that there was a need to document the effect of race on height as well as the interaction of race and age (18). In the present study, the equations derived from Philippine data were found to predict a more precise height approximation than models developed for Caucasians and Japanese-Americans.
CONCLUSION AND RECOMMENDATION
The specific equations developed using knee height and arm span can be used to estimate the height of Filipino elderly. Arm span and knee height are reasonable proxy indicators for height when it is difficult or not possible to measure height directly.
As this study utilized Filipino elderly 60 years old and over, it is therefore recommended that the formula be validated for use with other age groups and possibly other ethnic groups in the Philippines for greater utility value of these proxy indicators for height.
