Clinical Biomechanics 18 (2003) 207213 www.elsevier.

com/locate/clinbiomech

A comparison of the gaits of Chinese and Caucasian women with particular reference to their heelstrike transients
Wen-Ling Chen
a

a,*

, John J. OConnor b, Eric L. Radin

Department of Physical Therapy, National Cheng Kung University, No. 1, Ta-Hsueh Rd. Tainan 701, Taiwan, ROC b Oxford Orthopaedic Engineering Centre, Nueld Orthopaedic Centre, Headington, Oxford OX3 7LO, UK c Department of Orthopaedic Surgery, Tufts University School of Medicine, Box 561, Marion, MA 02738, USA Received 23 November 2001; accepted 28 November 2002

Abstract Objective. To examine the hypothesis that the dierences reported in incidence of osteoarthrosis in Chinese and Caucasians could be associated with dierences in habitual gait. Design. The eects of race and age on walking speed and heelstrike transient were examined. Background. The relatively low incidence of gonarthrosis in Chinese populations compared to Caucasians remains unexplained. Repetitive impulsive loading exhibited at heel strike in the walking process has been linked to the development of gonarthrosis, while the gait characteristics of people at dierent risk levels for gonarthrosis have not been compared quantitatively. Methods. The gait of 117 healthy women, 76 Chinese and 41 Caucasians, was studied with an optometric system and two force plates in an 8-m walkway. Natural walking speed, stride length, cadence and maximum loading rate at heelstrike were collected. Results. The Caucasian women over age 45 walked signicantly faster with signicantly higher maximal loading rate than agematched Chinese women (P < 0:005). Age eects on most gait parameters measured were found signicant in the Chinese group (P < 0:01) but not in the Caucasian group. Conclusions. Chinese women slow down their walking speed and reduce the cadence of their gait earlier in their life span and, thus, lower their heelstrike transients. Signicant racial dierences in gait might explain the lower prevalence of gonarthrosis reported in Chinese women. Relevance Signicantly larger heelstrike transients and signicantly faster walking speed were seen in the population at higher risk for gonarthrosis. Walking slowly with lower heelstrike transients might be an eective preventive measure against gonarthrosis. 2003 Elsevier Science Ltd. All rights reserved.
Keywords: Heelstrike transients; Gait; Gonarthrosis; Chinese; Caucasians

1. Introduction The relatively low incidence of osteoarthrosis (OA) of the knee, gonarthrosis, in Chinese compared to Westerners is intriguing and as yet unexplained (Hoaglund et al., 1973). Animal studies have shown that repetitive impulsive loading, applied suddenly, leads to subchondral microfracture, subsequent tidemark advance and progressive arthrotic changes (Sokolo, 1993; Radin et al., 1984). Cartilage failure has been hypothesized to result from the subsequent thinning
*

Corresponding author. E-mail address: wlc58@mail.ncku.edu.tw (W.-L. Chen).

of the overlying articular cartilage in this process (Kelly and OConnor, 1996; Burr and Schaer, 1997). Rapidly applied compressive loads can, theoretically, induce tensile stress both at the cartilage surface and at the cartilage/bone interface which could lead to tensile failure of the collagen network (Kelly and OConnor, 1996). For these reasons we have focused our attention on the slope of the heelstrike transient (HST), the peak dynamic force that occurs at the beginning of the stance phase of human gait when the heel rst contacts the ground. It has been shown that, in about 1/3 of human subjects, the rise time of this force is short enough to create impulsive loading (Radin et al., 1986).

0268-0033/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(02)00187-0

208

W.-L. Chen et al. / Clinical Biomechanics 18 (2003) 207213

There is circumstantial evidence to support the hypothesis that repetitive impulsive loading is correlated with knee OA. In 8090% of uni-compartmental osteoarthrotic subjects, full thickness cartilage erosion occurs initially on the anterior aspect of the tibial plateau and on the distal aspect of the femoral condyle, while the cartilage on the back of the plateau and condyle is still full thickness (White et al., 1991). These sites of the early arthrotic lesions are those in contact during the stance phase of gait so that it would be possible to associate their appearance with some vague form of wear and tear. This would not explain why only a proportion of the population is at risk of OA. The sites perfectly match the surfaces in contact during heelstrike (HS), at the beginning of stance phase (OConnor et al., 1989; Perry, 1992). At HS, some people exhibit a very high rate of loading with a distinct impulsive peak (Radin et al., 1986). These rapidly applied loads, generally ranging from 0.5 to 1.25 times body weight (BW) and occurring in only 525 ms (Simon et al., 1981), might cause microscopic damage if they are too fast to allow the interstitial uid to move. Although it is not possible to correlate present loading rates with the future development of OA directly in particular subjects, there is evidence suggesting the likelihood of such an eventuality. Subjects with pre-clinical knee pain were found to walk with a signicantly higher loading rate at HS than their age-matched pain-free control subjects (Radin et al., 1991). It is our hypothesis that individuals exhibiting high loading rates at HS dene the group most at risk of idiopathic OA. Unfortunately, the prospective study required to test this hypothesis would have to follow the adult life spans of groups of subjects and controls, an impractical proposition. However, gonarthrosis is known to have age, gender and racial predilections. For those under age 45, gonarthrosis is more common in men than in women, while after age 45, a steep rise in prevalence has been found in women (Lawrence et al., 1966; Hernborg and Nilsson, 1991; National Center for Health Statistics, 1979; Valkenburg, 1980). The occurrence of OA in Chinese compared to Caucasian populations has been shown to dier markedly and this dierence appears to involve both hips and knees (Hoaglund et al., 1973; Ota, 1979; Chou et al., 1994). In a random study, the prevalence of gonarthrosis in those aged 5564 was reported to be 7% for Chinese women, as opposed to 15% for Caucasian women (Hoaglund et al., 1973; Kellgen and Lawrence, 1958). Others have subsequently supported this nding (Chou et al., 1994). These ndings lead to a practical hypothesis: Caucasians, particularly older women who are at higher risk of gonarthrosis, walk in a dierent way from age-matched Chinese women. In particular, Caucasian women will walk with a signicantly higher HST than their age-matched Chinese controls. The implication of these ndings could be the basis for a treatment

of gonarthrosis in the patients with a high HST, since it is possible to teach people to walk dierently (Tashman and Radin, unpublished data).

2. Methods This study involved analysis of the natural gait of healthy Chinese and Caucasian women, including measurement of HSTs. One hundred seventeen healthy and asymptomatic women, 76 Chinese and 41 Caucasians, without any history of lower extremity symptoms or obvious pathology, were randomly recruited for gait analysis. Since the study took place in England, the Chinese subjects were recruited from immigrants working there under the assumption that they represented a reasonable cross-section of the general Chinese population. The Caucasian subjects were recruited from General Practitioners lists and the Chinese subjects were recruited from the Oxford neighbourhood. Ethics Committee approval was obtained for the study. All subjects were between the ages of 18 and 74 and divided into four groups: Caucasian women under 44, those over 45; Chinese women under 44; those over 45. They were age-matched and no signicant age dierences were found between the dierent younger and older racial groups (Table 1). Informed consent following the Royal College of Physicians guidelines for research involving human subjects was obtained before data collection. The gait analysis was carried out in the Oxford Orthopaedic Engineering Centre. A VICON optometric system (VICON370, Oxford Metrics Ltd., Oxford, England) and force plates (OR6-6-1000, AMTI, Newton, MA, USA) were used to collect kinematic and kinetic data for calculating walking speed and the HST during walking. In this study, the video signals were collected at a rate of 50 Hz and the analogue signals from the force plate were digitized at a sampling rate of 1000 Hz. To
Table 1 Mean (SD) of age (year) and anthropometric data including height (cm), BW (kg), and BMI (BW divided by the square of height, expressed in kg/m2 ) Younger group (age 1844) Caucasians (n 18) Age (year) Heighta (cm) Weighta (kg) BMIb (kg/m2 ) 29 (7) 166.6 (5.7) 66.4 (9.6) 23.9 (3.2) Chinese (n 45) 32 (6) 158.0 (4.6) 54.6 (6.4) 21.9 (2.7) Older group (age 4574) Caucasians (n 23) 55 (9) 162.7 (7.1) 70.0 (13.0) 26.4 (4.5) Chinese (n 31) 57 (7) 155.9 (5.6) 59.6 (8.9) 24.6 (4.2)

a Signicant racial dierences exist in the younger groups and in the older groups, by Student t tests (P < 0:005). b Signicant racial dierences exist only in the younger groups by Student t tests (P < 0:05).

W.-L. Chen et al. / Clinical Biomechanics 18 (2003) 207213

209

quantify the HST, a quadratic polynomial was tted to the trace of the vertical component of ground reaction force immediately after HS before the rst impulsive peak was reached and the maximum slope during the rise time of the HST (LRmx , maximum loading rate) was calculated (see Fig. 1) (Gill, 1996). This variable was preferred to the estimation from the digitized data of the magnitude of the peak of the rapidly varying force. At a collection rate of 1000 Hz, we had only a limited number of data points (within 530 ms), adequate to determine the slope but nowhere near adequate to give a good estimate of the peak value. To reduce the inuence resulting from anthropomorphic features, LRmx was then normalized by individual BW and expressed in BW/s. Since loading rate and walking speed have been found to be correlated (Collins and Whittle, 1989), walking speed was also calculated by using the spatial data obtained from the ankle markers between the HST where LRmx was collected and its consecutive HST. Additional markers applied on other landmarks in the right leg were used for further kinematic interpretation regarding the motor control of HSTs for future study. The subjects were dressed in shorts and instrumented with reective markers (25 mm in diameter) on both lateral malleoli. Before the measurements were taken, the subjects were asked to practice the walking procedure along an 8-m walkway until they could reproduce what they considered to be their natural walking speeds. To avoid alterations in natural walking pattern during the measurements, the subjects were not given specic instructions to step on the force plate. At least 10 successful walking trials, which included complete force data, were collected for each subject, ve trials for each
(a)

1200
F1

400 300
***

F0 fitted curve Fz1 data points A

1000

200

Fz1(N)
100 800 0
HS 2.94

Fz1(N )
600

(b)
4 400
F0

2.95

2.96

Time(s)
x 10 4
A

leg respectively. All trials were performed barefoot, since variations in the foot wear worn by individual subjects could lead to marked dierences in the results obtained (Simon et al., 1981; Forner et al., 1995; Jorgensen and Bojsen Moller, 1989; Jorgensen and Ekstrand, 1988). In the circumstances, the true dierences in HST between dierent populations would be detected more easily without interference from the material of shoe insert, heel pads, or even heel counter. Besides, it was reported that those who exhibited higher than average loading rates without shoes also did so with shoes (Yang and Radin, personal communication). Averages from the successful walking trials for each leg were used in all analyses. Bilateral comparisons to test left/right symmetry in walking speed and LRmx were examined rst by using a paired t test. Data from right leg trials were then chosen for further analysis, since no signicant dierences were found between both leg trials (step speed, right: 1.24(0.23) m/s, left: 1.24(0.23) m/s, t 0:55; LRmx , right: 71.9(37.6) BW/s, left: 70.4(35.7) BW/s, t 1:06). To examine the changes of natural walking speed across age span, a one-way A N O V A was used to test the dierences in natural walking speed among dierent age categories by 10-year increments. A two-way A N O V A was chosen to examine the overall main eects of age and race on natural walking speed. In the circumstances that the interaction between age and race factor was signicant, an investigation of simple eects for each factor was conducted by using Student t tests. Considering the steep rise of gonarthrosis prevalence reported in women after age 45, both Caucasian and Chinese groups were divided at age 45 to examine the age eects of gait changes after age 45. LRmx was found to have a nonnormal distribution for all the subgroups studied in the present experiment, as previously reported (Radin et al., 1991). MannWhitney tests were therefore used to examine the simple race and age eects on LRmx (Hinton, 1995). To eliminate the eects of anthropomorphic features, which diered signicantly between the racial groups (Table 1), the normalized LRmx , after adjusting for leg length and body mass index (BMI) (BW divided by the square of height), were also examined by the same design.

3 2 200

dFz/dt (N/s) 1
0

3. Results 3.1. Eects of age on walking speed

2.94 2.95 2.96

0 2.5

HS 3

3.5

-1

Time(s)

(c)

Time(s)

Fig. 1. Regression method of calculating maximal loading rate, based on the vertical component of ground reaction force (Fz) : (a) typical Fz trace against time, showing an impulse at F0 occurring between (b) the initial part of Fz between HS and F0 (star points) at an expanded scale and the tted polynomial regression line () (c) the rst derivative of the tted polynomial and the maximal value (at A).

Table 2 shows that natural walking speed decreased signicantly with age for the Chinese women, but not for the Caucasian women. The results of one-way A N O V A show no signicant dierences in walking speed among any age category in the Caucasian women. On the other hand, the results of multiple comparisons

210

W.-L. Chen et al. / Clinical Biomechanics 18 (2003) 207213 Table 3 Mean (SD) of natural walking speed (m/s) and other temporal-distance parameters (stride length, expressed in m; normalized stride, expressed in body height, BH; cadence, expressed in stride/min) in dierent ageracial groups Younger group (age 1844) Caucasian women (n 17) Speed (m/s) Stride length (m) Normalized stride (BH) Cadence (stride/min) 1.40 (0.17)b 1.37 (0.11)b 0.82 (0.01)b 61.5 (4.6) Chinesea women (n 44) 1.23 (0.17) 1.22 (0.12) 0.77 (0.01) 60.2 (4.8) Older group (age 4574) Caucasian women (n 23) 1.40 (0.18)b 1.34 (0.14)b 0.82 (0.01)b 62.8 (4.2)b Chinesea women (n 31) 1.02 (0.16) 1.08 (0.23) 0.69 (0.15) 55.5 (5.2)

Table 2 Mean (SD) of natural walking speed (m/s) in dierent age category Age subgroup (Sample sizeCaW=ChW) Age Age Age Age Age Age
a

Caucasian women (CaW) 1.47 1.40 1.38 1.41 1.51 1.29 (0.18) (0.39) (0.13) (0.16) (0.28) (0.07)

Chinese women (ChW) 1.37 1.26 1.15 1.05 1.05 0.86 (0.17) (0.17) (0.15) (0.18)a (0.15)b (0.14)c

1524 2534 3544 4554 5564 6574

year year year year year year

(3/5) (9/22) (5/17) (14/9) (4/18) (5/4)

Subjects in this group walk signicantly more slowly than age 1524 group (P < 0:05). b Subjects in this group walk signicantly more slowly than age 1524 group and age 2534 group (P < 0:05). c Subjects in this group walk signicantly more slowly than age 1524 group and age 2534 group (P < 0:05).

using Schee post hoc tests for the Chinese group revealed that all subgroups after age 45 walked signicantly more slowly than the age 1524 subgroup, and both subgroups after age 55 walked signicantly more slowly than the age 2534 group (P < 0:05). It appears that the point at which the Chinese women reduced walking speed signicantly was age 45. Two-way A N O V A of natural walking speed reveals signicant interaction between the age and race factors (P < 0:005). Student t tests were therefore conducted and conrmed that a signicant simple age eect existed in the Chinese group (P < 0:001) but not in the Caucasian group. Signicant age eects on stride length and cadence were also noted in the Chinese (stride length, P < 0:005; cadence, P < 0:001) but not in the Caucasian women (Table 3), even after adjusting for body height (normalized stride), P < 0:01. 3.2. Eects of age and race on HST measurements Interactions between age and race factors were also noted in HST measurements. Signicant HST dierences between the younger and the older groups were found in the Chinese women (P < 0:01, MannWhitney), but not in the Caucasian women. The Chinese women walked signicantly less heavily after age 45, but the Caucasian women did not show such signicant changes with increasing age (Table 4). This was true even when the data was adjusted for body height (Chinese: MannWhitney, P < 0:05). 3.3. Eects of race The results of MannWhitney tests showed that the Caucasians walked signicantly faster with signicantly higher HST than the Chinese, either in the younger (age 1844) or the older (age 4574) groups. (Speed for the younger subjects: P < 0:005, for the older: P < 0:001, see Table 3; LRmx , for the younger: P < 0:001, Mann

a Signicant simple age eects (comparisons between the younger and the older) on temporal-distance gait parameters were found in Chinese (speed, t 5:25, P < 0:001; stride, t 3:05, P < 0:005; normalized stride, t 2:76, P < 0:01; cadence, t 4:01, P < 0:001), but not in Caucasians. b Signicant racial dierences for all the temporal-distance measurements exist in the older (speed, t 8:02, P < 0:001; stride, t 5:0, P < 0:001; normalized stride, t 4:19, P < 0:01; cadence, t 5:74, P < 0:001) and the younger (speed, t 3:63, P < 0:005; stride, t 4:53, P < 0:001, normalized stride, t 2:80, P < 0:01), except for cadence in the younger group.

Table 4 Mean (SD) of maximal loading rate (LRmx , expressed in body weight per second, BW/s) and the data after adjusting for height and BMI in dierent age-racial groups Younger group (age 1844) Caucasians (n 18) LRmx b (BW/s) LRmx /heightb (BW/s/m) LRmx /BMIb (BW m2 /s/kg) 110.0 (51.2) 65.8 (30.3) 4.8 (2.7) Chinesea (n 45) 65.7 (27.4) 41.6 (17.6) 3.1 (1.4) Older group (age 4574) Caucasians (n 23) 82.6 (34.3) 50.7 (20.9) 3.3(1.6) Chinesea (n 31) 49.6 (24.7) 31.8 (15.6) 2.1 (1.2)

a Age eect (comparisons between the younger and the older) was found signicant on LRmx in Chinese women (MannWhitney, P < 0:01) but not in Caucasian women, even after adjusting for body height (LRmx /height: Chinese, MannWhitney, P < 0:05). b Signicant racial dierences of LRmx exist in both the younger and the older (the younger: MannWhitney, P < 0:001, the older: Mann Whitney, P < 0:001), Caucasian women walk more heavily with signicantly higher LRmx than Chinese women even after adjusting for body height (the younger: MannWhitney, P < 0:001, the older: MannWhitney, P < 0:001) or BMI (the younger: MannWhitney, P < 0:01, the older: MannWhitney, P < 0:005).

Whitney, for the older: P < 0:001, MannWhitney, see Table 4). This was true even when the data was adjusted for body height and BMI (Table 4). Signicant racial dierences for other temporaldistance measurements were also noted, except for

W.-L. Chen et al. / Clinical Biomechanics 18 (2003) 207213

211

cadence in the younger groups (Table 3). The Caucasian women walked with a signicantly longer stride than the Chinese women (stride length, the older group: P < 0:001, the younger group: P < 0:001) even after adjusted for body height (normalized stride, P < 0:001 for the older group, P < 0:01 for the younger group), but racial dierences in cadence were only found in the older group (P < 0:001).

4. Discussion The causation of OA is multifactorial, involving factors such as joint structure, alignment, development. The present study is concerned with idiopathic OA with no obvious explanation such as post-traumatic or post-inammation disease. It has been suggested that repetitive rapidly applied impulsive loading creates osteoarthrotic changes (Radin et al., 1985). It is believed that such loading creates subchondral micro-fractures and that this cumulative damage reactivates the secondary center of ossication, leading to progressive articular cartilage loss (Radin et al., 1984). This suggests an underlying relationship between OA and age, borne out by clinical observation. Alterations in subchondral bone and the increase of bone mass were found only in a 50-ms repetitive loading protocol (1 Hz) rather than a 500-ms protocol in an animal study (Farkas et al., 1987). The duration of the HST in human walking was reported to be 525 ms (Simon et al., 1981). The magnitude of the peak force varied from 0.5 to 1.25 times BW and its frequency components from 10 to 75 Hz, measured by force plates with a high resonant frequency of 1300 Hz (Simon et al., 1981). At HS, some people exhibit a very high rate of loading with a distinct impulsive peak (Radin et al., 1986). Subjects with pre-clinical knee pain were found to walk with a signicantly higher loading rate at HS than their age-matched, pain-free control subjects (Radin et al., 1991). This evidence suggested the hypothesis that subjects exhibiting high HST may form the group at risk of gonarthrosis. The present study shows that the older Chinese women subjects walked more slowly and with a lower loading rate than the younger Chinese women. The Caucasian women walked signicantly faster and more heavily with signicantly higher LRmx at HS than the Chinese women, particularly after age 45. These dierences are consistent with the lower prevalence of gonarthrosis reported in the Southern Chinese population (prevalence of gonarthrosis, age 5564: Chinese women, 7%; Caucasian women, 15%) (Hoaglund et al., 1973). The mechanism for Chinese women to slow down their walking with age might result from the shorter stride and, especially, from the reduced cadence over age. As seen in Table 3, signicant age eects on all the tem-

poral-distance parameters were found in the Chinese but not in the Caucasian subjects. In addition, before age 45, no signicant racial dierence was found in cadence while, after age 45, racial dierences in stride length, cadence and walking speed became signicant. From similar studies comparing Chinese and Caucasian men, the eect of age on walking pattern (speed and HST) was more prominent in men than in women (Chen, 1998). Older men of both races were reported to walk signicantly more slowly and less heavily than younger men, an eect not seen in the older Caucasian women. On the other hand, the gender eect was more signicant in young than in old people, without signicant interaction from the racial factor (Chen, 1998). Young men walked signicantly faster and more heavily than young women but the dierences between older men and older women were not signicant. People more at risk for gonarthrosis, such as women in comparison to men, seem to be those who keep walking with higher HST until very old (Chen, 1998). The prevalence of hip OA has been reported to be very low among Chinese in Hong Kong (Hoaglund et al., 1973), in Beijing (Nevitt et al., 2002), among elderly persons in Japan (Shichikawa et al., 1966), and among Japanese and Chinese living in Hawaii (Oishi et al., 1998) and San Francisco (Hoaglund et al., 1995). If hip OA is uncommon in Chinese and Japanese populations, this might reect a low occurrence of OA in general. However, this view has recently been questioned. Inoue et al. studied adult skeletons excavated from archaeological sites in Japan, China and France, which were assessed for OA by the presence of eburnation (Inoue et al., 2001). Their ndings suggested the Asian skeleton knees had a higher prevalence of tibiofemoral joint OA (Inoue et al., 2001). However none of the Caucasian skeletons investigated was reported to have tibiofemoral joint OA. Considering the known prevalence of knee OA in Caucasians today (1535%), it is questionable whether conclusions drawn from this paleopathological paper are representative of modern Caucasian populations. Another recent study of a subgoup of Beijing women found them to have a higher incidence of lateral compartment knee OA than Caucasians. Gonarthrosis of this type, rare in Caucasian populations, was more prevalent in this subgroup of elderly Northern Chinese than in white women (Felson et al., 2002). However, this Northern subgroup may not be representative of Southern Chinese. In our study the Chinese subjects were immigrants from Southern China working in England who shared for a period the same environment with the Caucasians investigated. In the populations we studied our data suggest a very clear racial correlation between HST and walking speed with the clinical incidence of gonarthrosis. Chinese women tend to slow down their walking speed earlier in

212

W.-L. Chen et al. / Clinical Biomechanics 18 (2003) 207213 Chapman, K., Mustafa, Z., Irven, C., Carr, A.J., Clipsham, K., Smith, A., Chitnavis, J., Sinsheimer, J.S., Bloomeld, V.A., McCartney, M., Cox, O., Cardon, L.R., Sykes, B., Loughlin, J., 1999. Osteoarthritis-susceptibility locus on chromosome 11q, detected by linkage. American Journal of Human Genetics 65, 167174. Chen, W.-L., 1998. Impulsive Loading in Gonarthrosis. D.Phil. Thesis. University of Oxford, Oxford. Chou, C.T., Pei, L., Chang, D.M., Lee, C.F., Schumacher, H.R., Liang, M.H., 1994. Prevalence of rheumatic diseases in Taiwan: a population study of urban, suburban, rural dierences. The Journal of Rheumatology 21, 302306. Collins, J.J., Whittle, M.W., 1989. Inuence of gait parameters on the loading of the lower limb. Journal of Biomedical Engineering 11, 409412. Farkas, T., Boyd, R.D., Schaer, M.B., Radin, E.L., Burr, D.B., 1987. Early vascular changes in rabbit subchondral bone after repetitive impulsive loading. Clinical Orthopaedics and Related Research 219, 259267. Felson, D.T., Naimark, A., Anderson, J., Kazis, L., Castelli, W., Meenan, R.F., 1987. The prevalence of knee osteoarthritis in the elderly: the Framingham osteoarthritis study. Arthritis and Rheumatism 30, 914918. Felson, D.T., Nevitt, M.C., Zhang, Y., Aliabadi, P., Baumer, B., Gale, D., Li, W., Yu, W., Xu, L., 2002. High prevalence of lateral knee osteoarthritis in Beijing women compared with Framingham Caucasian subjects. Arthritis and Rheumatism 46, 12171222. Forner, A., Garcia, A.C., Alcantara, E., Ramiro, J., Hoyos, J.V., Vera, P., 1995. Properties of shoe insert materials related to shock wave transmission during gait. Foot and Ankle International 16, 778 786. Gill, H.S., 1996. The Mechanics of Heelstrike During Level Walking. D.Phil. Thesis. University of Oxford, Oxford. Hernborg, J.S., Nilsson, B.E., 1991. Age and sex incidence of osteophytes in the knee joint. Acta Orthopaedica Scandinavica 44, 6668. Hinton, P.R., 1995. Two sample nonparametric analysis. In: Hinton, P.R. (Ed.), Statistics Explained. Routledge, London, pp. 209224. Hoaglund, F.T., Yau, A.C.M.C., Wong, W.L., 1973. Osteoarthritis of the hip and other joints in Southern Chinese in Hong Kong. Journal of Bone Joint Surgery A 55, 545557. Hoaglund, F.T., Oishi, C.S., Gialamas, G.G., 1995. Extreme variations in racial rates of total hip arthroplasty for primary coxarthrosis: a population-based study in San Francisco. Annals of Rheumatic Disease 54, 107110. Inoue, K., Hukuda, P., Fardellon, P., Yang, Z.Q., Nakai, M., Katayama, K., Ushiyama, T., Saruhashi, Y., Huang, J., Mayeda, A., Catteddu, I., Obry, C., 2001. Prevalence of large-joint osteoarthritis in Asian and Caucasian skeletal populations. Rheumatology 40, 7073. Jorgensen, U., Bojsen Moller, F., 1989. Shock absorbency of factors in the shoe/heel interactionwith special focus on role of the heel pad. Foot and Ankle 9, 294299. Jorgensen, U., Ekstrand, J., 1988. Signicance of heel pad connement for the shock absorption at heel strike. International Journal of Sports Medicine 9, 468473. Kellgen, J.H., Lawrence, J.S., 1958. Osteo-arthrosis and disk degeneration in an urban population. Annals of Rheumatic Disease 17, 388397. Kelly, P.A., OConnor, J.J., 1996. Transmission of rapidly applied loads through cracked articular cartilage. Proceedings of the Institution of Mechanical Engineers. Part H, Journal of Engineering in Medicine 210, 2749. Lawrence, J.S., Bremner, J.M., Bier, F., 1966. Osteoarthrosis: prevalence in the population and relationship between symptoms and X-ray changes. Annals of Rheumatic Disease 25, 124. Loughlin, J., Mustafa, Z., Smith, A., Irven, C., Carr, A.J., Clipsham, K., Chitnavis, J., Bloomeld, V.A., McCartney, M., Cox, O.,

their life span than Caucasian women. This signicant racial eect, in both walking speed and HST, might explain the lower prevalence of gonarthrosis reported in Chinese populations. In addition, our studies of Chinese and Caucasian men (Chen, 1998), who have a similar incidence of gonarthrosis, demonstrated a slowing of cadence and a diminution of HST with age in both groups. The fact that this age eect was signicant in men but not in women might further explain the steep rise in gonarthrosis in Caucasian women after age 45. The way people walk could be merely habit but it is probable that much of our walking pattern is inherited. A twins study reported signicantly greater intraclass correlations in monozygotic pairs (n 95) than in dizygotic pairs (n 92) for usual walking speed measurements (Felson et al., 1987). There is recent evidence of genetic linkage with the incidence of OA (Carmelli et al., 2000; Chapman et al., 1999; Loughlin et al., 2000). The authors attributed the linkage to the chromosome locus that could inuence the mass of the subchondral bone under the articulating cartilage as a regulator. However, the increased subchondral bone could be the result of repeated impulsive loading at HS (Farkas et al., 1987). Gait patterns may well be genetically determined. Whether genetic or environmental inuence is dominant in gait control, the present study has demonstrated signicantly faster walking speed with signicantly higher HST in Caucasian women, a group at higher risk for gonarthrosis. The easiest way to reduce loading rate is to slow down (Chen, 1998). This can be supported by the walking patterns of the older Chinese women reported in the present study. Walking slowly with lower HST might be an eective preventive measure against OA of the knee. We do not mean to imply that repetitive impulsive loading is the sole cause or even a contributory factor in all patients with OA, but it is possible that it is a factor in many cases. Certainly, there is enough suggestive evidence in its favor to give gait alteration a good clinical trial. Acknowledgement Dr. Chen was awarded a government scholarship by the Ministry of Education in Taiwan, ROC.

References
Burr, D.B., Schaer, M.B., 1997. The involvement of subchondral mineralized tissues in osteoarthrosis: quantitative microscopic evidence. Microscopy Research and Technique 37, 343357. Carmelli, D., Kelly-Hayes, M., Wolf, P.A., Swan, G.E., Jack, L.M., Reed, T., Guralnik, J.M., 2000. The contribution of genetic inuences to measures of lower-extremity function in older male twins. Journal of Gerontology, Biological Sciences A 55, B49 B53.

W.-L. Chen et al. / Clinical Biomechanics 18 (2003) 207213 Sinsheimer, J.S., Sykes, B., Chapman, K.E., 2000. Linkage analysis of chromosome 2q in osteoarthritis. Rheumatology 39, 377381. National Center for Health Statistics. Basic data on arthritis knee, hip and sacroiliac joints, in adult ages 2574 years, United States 1971 1975. National Center for health statistics, Rockville, MD, 1979 (Vital and health statistics. Series 11, no. 213) (DHEW publication no. (PHS) 79-1661). Nevitt, M.C., Xu, L., Zhang, Y., Lui, L.Y., Yu, W., Lane, N.E., Qin, M., Hochberg, M.C., Cummings, S.R., Felson, D.T., 2002. Very low prevbalence of hip ostearthrosis among Chinese elderly in Beijing, China, compared with whites in the United States: the Beijing osteoarthritis study. Arthritis and Rheumatism 46, 1773 1779. OConnor, J.J., Shercli, T.L., Biden, E., Goodfellow, J.W., 1989. The geometry of the knee in the sagittal plane. Proceedings of the Institution of Mechanical Engineers. Part H, Journal of Engineering in Medicine 203, 223233. Oishi, C.S., Hoaglund, F.T., Gordon, L., Ross, P.D., 1998. Total hip replacement rates are higher among Caucasians than Asians in Hawaii. Clinical Orthopaedics and Related Research 353, 166 170. Ota, H., 1979. Prevalence of osteoarthrosis of the hip and other joints in Japanese population. Nippon Seikeigeka Gakkai Zasshi Journal of the Japanese Orthopaedic Association 53, 165180. Perry, J., 1992. Gait Analysis: Normal and Pathological Function. Slack Inc., Thorofare, NJ. Radin, E.L., Martin, R.B., Burr, D.B., Caterson, B., Boyd, R.D., Goodwin, C., 1984. Eects of mechanical loading on the

213

tissues of the rabbit knee. Journal of Orthopaedic Research 2, 221 234. Radin, E.L., Martin, R.B., Burr, D.B., Caterson, B., Boyd, R.D., Goodwin, C., 1985. Mechanical factors inuencing cartilage damage. CIBA-Geigy, Paris, France. Radin, E.L., Whittle, M.W., Yang, K.H., Jeerson, R., Rodgers, M.M., Kish, V.L., OConnor, J.J., 1986. The heelstrike transient, its relationship with the angular velocity of the shank, and eects of quadriceps paralysis. In: Lantz, S.A., King, A.I. (Eds.), Advances in Bioengineering. American Society of Mechanical Engineering, New York, pp. 121123. Radin, E.L., Yang, K.H., Riegger, C., Kish, V.L., OConnor, J.J., 1991. Relationship between lower limb dynamics and knee joint pain. Journal of Orthopaedic Research 9, 398405. Shichikawa, K., Mayeda, A., Komatsubara, Y., Yamamoto, T., Akabori, O., Hongo, I., 1966. Rheumatic complaints in urban and rural populations in Osaka. Annals of Rheumatic Disease 2, 2531. Simon, S.R., Paul, I.L., Mansour, J., Munro, M., Abernethy, P.J., Radin, E.L., 1981. Peak dynamic force in human gait. Journal of Biomechanics 14, 817822. Sokolo, L., 1993. Microcracks in the calcied layer of articular cartilage. Archives of Pathology and Laboratory Medicine 117, 191195. Valkenburg, H.A., 1980. Clinical versus radiological osteoarthritis in general population. In: Peyron, J.G. (Ed.), Epidemiologie de larthrose. Geigy, Paris, pp. 5358. White, S.H., Ludkowski, P.F., Goodfellow, J.W., 1991. Anteromedial osteoarthritis of the knee. Journal of Bone Joint Surgery [B] 73, 582586.