Bone, 11,211-214 (1990) Printed in the USA. All rights reserved.

8756-3282190 $3.00 + .OO Copyright 0 1990 Pergamon Press plc

Wrist Fracture, Heel Bone Density and Thoracic Kyphosis: A Case Control Study R. W. PORTER, K. JOHNSON and J. D. S. MCCUTCHAN

Department of Orthopaedics, Doncaster Royal Infirmary, Doncaster, England

Address for correwondence and reprints: R. W. Porter, M.D., Department of Orthopaedics, Doncaster Royal Infirmary, Armthorpe Road, boncaster, England.

Abstract

The heel bone density measured by Broadband Ultrasound Attenuation (BUA), the thoracic kyphosis measured by a Kyphometer, height, and weight were compared between 294 women over 49 years of age who sustained a wrist fracture and 294 age-matched women who had not had previous wrist, hip or spine fracture. The BUA was sigulfkantly less in the women who had wrist fracture @ < 0.0005), though there was a considerable overlap between the two popula- tions. The women with wrist fracture had significantly greater thoracic kyphosis @ < 0.0005) and smaller stature @ < 0.0005). There was no siguificaut difference ln weight. There was a significant tendency (p < 0.0005) for women in the fracture patient group to have both poor BUA and greater kyphosis.

Key Words: Osteoporosis-Wrist fracture-Kyphosis-Ul- trasound.

Introduction

Women with fracture of the wrist have a significant reduc- tion in forearm bone mineral density (Consensus Confer- ence: Osteoporosis 1984; Hesp et al. 1984; Baran et al. 1988). They are at risk of sustaining a later hip fracture (Krolner et al. 1982; Owen et al. 1982; Horowitz et al. 1988; Lauritzen et al. 1988; ) but the degree of lower limb osteoporosis at the time of fracture is unknown (Krolner et al. 1982). We have measured the heel bone density by ul- trasound and also the degree of kyphosis in women sus- taining a wrist fracture, comparing them with age-matched controls.

Patients, Methods and Results

Women with a fracture of the distal radius who presented at Doncaster Royal Infirmary between October 1985 and January 1988, who were over 49 years of age, were invited to participate in the study. They were compared with vol- unteer control women attending general practice surgeries and orthopaedic clinics for unrelated conditions who were unaware of previous wrist, hip or spine fractures.

The heel bone density was measured by Broadband Ul- trasound Attenuation (BUA), (Langton et al. 1984) using a Walker-Sonix Ultrasound Bone Analyser 1001 with two 1MHz ultrasound transducers operating in the frequency range 0.2 to 0.8 MHz. The BUA technique compares fa- vorably with dual photon absorptiometry of the vertebrae (Baran et al. 1988) and other techniques (Poll et al. 1986; Evans et al. 1987; Petley et al. 1987).

The foot was placed in the water tank with sufficient shims beneath and behind the heel to ensure that the body of the OS calcis was between the two transducers. The number of shims was determined from circumferential hindfoot and forefoot measurements. Ultrasound attenua- tion was then recorded between the two transducers, and the heel measurement obtained after subtracting a previous water trace.

It can take a few minutes for the Analyzer to settle as air is displaced from the surface of the skin by detergent in the water. In order to ensure that results were not recorded until the system had stabilized, the following procedure was adopted. Serial measurements of the gradient of ultra- sound attenuation across the frequency range 0.2 to 0.8 MHz were obtained until three successive values were ob- tained within 0.5 dB/MHz. The third value was noted. The foot was removed from the water tank and then replaced, obtaining measurements again until three more readings were within 0.5 dB/MHz. The third of these was again noted. This cycle was repeated until two successive noted values were within 5 dB/MHz. The mean of these two values was recorded as the BUA for that heel.

Thoracic kyphosis between Dl and D12 was measured in degrees with a Kyphometer. The repeatability of this device has been reported previously (Salisbury and Porter 1987). They obtained a mean difference of measurement of 5.53”, standard deviation 3.02”, when 17 subjects were measured on two separate days. For the purposes of the present study, subjects were asked to stand while verte- brae Dl and D12 were identified by palpation. One heel of the Kyphometer was placed on the dorsal spine at Dl and the Kyphometer opened until the second heel rested against D12. The angle in degrees between these two spinous processes could then be recorded from the scale on the Kyphometer. The standing height and weight were also recorded.

During the period of study there were 47,500 women residing in the catchment area of Doncaster Royal Infii-

211

212 R. W. Porter et al.: Wrist fracture, bone density and kyphosis

mary who were over 49 years of age. Five hundred sixty- one women sustained a fracture of the wrist. They had a median age of 65 years, with a range of 48 years (50-98). The fracture rate in each five-year age-band is shown in Fig. 1. Three hundred seventy-seven nonfracture control women were examined, and 294 of these age-matched by age in completed years with 294 fracture patients.

A statistical comparison was made between the Broad- band Ultrasound Attenuation, the thoracic kyphosis, standing height and weight in the two populations. The data approximated a normal distribution with similar stan- dard deviations in the two populations and was therefore assessed by Student’s two-sample t test. This test was used because it implicates differences between every age- matched pair of subjects in its result.

Firstly, the reproducibility of the technique during the present work was assessed by using data from 66 patients who had BUA measurements on two different days. A mean difference of 5.8 dB/MHz was obtained (standard de- viation 5.2 dB/MHz, standard error 0.64 dB/MHz). The precision of the technique is 0.5 dB/MHz.

The BUA measurement for the two populations is shown in Fig. 2. There was a considerable overlap between the two populations. The ultrasound attenuation was, how- ever, significantly less (p < 0.0005) in the patients with wrist fracture than in the controls (mean 65.5 dB/MHz, standard deviation 18.5 dB/MHz, compared with mean 78.0 dB/MHz, standard deviation 22.5 dB/MHz, respec- tively). They also had significantly greater (p < 0.0005) thoracic kyphosis (mean 42”, standard deviation lo”, com- pared with mean 37”, standard deviation 9”, respectively) and smaller @ < 0.0005) stature (mean 157 cm, standard deviation 6 cm, compared with mean 160 cm, standard de- viation 6 cm, respectively). There was no significant differ- ence in weight.

The BUA for combined fracture subjects and controls had a correlation of r = 0.41 (p < 0.001) with age (fracture cases alone r = 0.42, p < 0.001; control subjects alone r = 0.46, p < 0.001). A comparison of the kyphosis for both populations with age showed no correlation in either group when considered individually, but when the two groups were combined, a correlation coefficient of r = 0.23 (p < 0.0005) was observed. There was a significant negative correlation between BUA and thoracic kyphosis (r = 0.25, 0.02 > p > 0.01) for the fracture group (Fig. 3), but no such relationship in the control group (r = 0.059; p > 0.1).

Evidence was sought for the co-existence of osteopo-

'01 I I

Age group (year)

Fig. 1. Incidence of Colles’ fracture in women over 49 years of age in Doncaster.

150

1

p< 0.0005

140 1

.: .: j;.. -,:: -----

. . :. /:.. : i.::.. j:;:;.. j::. ::

ii:‘. . .

/:

:i:.

I:: :..

: j/i.. /:, ::: //. :..: /ij: ;jj:. 1.:: i:.. j:.

i.. /in.______

.I.. :: ::.,

j;:

j:. /jj t:. I;_ jji

::

Fig. 2. Broadband Ultrasound Attenuation of the OS calcis in 294 women with fracture of the wrist and 294 controls matched for age. Broken lines indicate mean and two standard deviation values in the control women.

rotic indices in the OS calcis and thoracic spines. The dis- tribution of both fracture patients and controls displaying evidence of osteoporosis at both these anatomical sites was studied. The 50th. centile value for BUA and kyphosis of the control group was taken as an arbitrary cut-off value. Values lower than the cut-off were defined as “por- otic” and those greater as “healthy.” Individuals were identified where both values were on the same side, and where they were on opposite sides of the cut-off (Table I). A test for homogeneity of proportions between the fracture and control groups (A vs. B vs. C) indicated unequal pro- portions (p < 0.0005; chi-squared = 112.536; 2” of freedom).

Partitioning the overall GL into a further test of interest, comparing the “porotic” group with all other cases (A vs. B + C!), G2 was 105.914 (1” of freedom) (p < 0.0005).

Discussion

Ultrasound attenuation of the OS calcis is a relatively simple technique of bone measurement; being safe and ra-

R. W. Porter et al.: Wrist fracture, bone density and kyphosis 213

80 -

Kyphosis (degrees)

60 -

40 -

20 -

o-

+ +

+

+ +

+ + r = 0.25 0.02 > p > 0.01

+ + +

+ +++ :

+ ++ + + + +* +-I+

+ + ++ ?++t#t+ +

+ +++ +#t++ti $+ + + f+*+ +$ #+t++l;” +

+ ++ &$%l ?+I ++++ +

+ + +I-+++ ++ +#-n+A ++ +f + ++

-I#+ + A+ + +y j+ +

+ ; #+++ ++

+ + ++ -I+

+

+

c 1 I I I

40 80 120

B.U.A. (dB/MHz)

Fig. 3. The thoracic kyphosis (measured by Kyphometer) and Broadband Ultrasound Attenuation of the OS calcis in 294 women with fracture of the wrist.

diation-free it was ideal for this large population study. Ul- trasound differs from other methods of measurement, probably being related to the bone’s architecture as well as the mineral content, both important parameters contrib- uting to the overall strength of the bone. It is of interest that there was a significant difference of BUA between the fracture and the nonfracture populations, particularly as the measurements were recorded from the heel. It sup- ports the view of Wasnich et al. (1985) that the OS calcis is a good site for assessment of the risk of nonspinal osteopo- rotic fractures.

The degree of overlap between the fracture and the nonfracture populations precludes the use of Colles’ frac- ture as a screening test for subsequent hip fracture, but the principle of selecting women with wrist fracture for pro- phylactic screening against osteoporotic hip fracture (Con-

sensus Conference: Osteoporosis 1984) appears to be sound.

Krolner et al. (1982) reported a reduced bone mass in the axial skeleton in women with Colles’ fracture, but we believe that this is the first time that thoracic kyphosis has been assessed. The increased kyphosis implies that women with wrist fracture have already sustained multiple minor micro-fractures with vertebral wedging, and that they are at risk of sustaining further vertebral compression frac- tures .

Examination of the co-existence of osteoporotic indices at two anatomical sites suggests that the population of women who sustained a wrist fracture were significantly different to the control population. The greater proportion of fracture patients who were found to have both poor BUA values in their heel bones and greater thoracic ky-

Table I. Comparison between the individuals with both BUA and kyphosis “porotic,” “ healthy,” sides of the watershed-for fracture patients and controls.

and those with one value on opposite

Fracture group

Control group

Both values “porotic” (A)

frequency %

156 53.1

42 14.3

One value each “porotic”/“healthy”

(B)

frequency %

114 38.8

180 61.2

Both values “healthy” (C)

frequency %

24 8.1

72 24.5

214 R. W. Porter et al.: Wrist fracture, bone density and kyphosis

phosis suggests that osteoporosis in the fracture patient group was generalized.

The small stature of our population of women is at vari- ance with previous observations (Nilsson and Westlin 1974). The reduced stature probably results from the tho- racic kyphosis, rather than preceeds it.

Acknowledgments: We would like to thank Sanofi, UK, for their financial support, Mrs. Jean Reynolds for her secretarial assis- tance, and Mr. Carry Swann for the illustrations.

References

Baran, D. T.; Kelly, A. M.; Karellas, A. Ultrasound attenuation of the OS calcis in women with osteoporosis and hip fractures. C&if. Tissue In?. 138-42; 1988.

Consensus Conference: Osteoporosis. J. Am. Med. Assoc. 799-802; 1984. Evans, W. D.; Crawley, E. 0.; Compston, J. E.; Evans, C.; Owen, G. M.

A comparison of broadband ultrasound attenuation with single photon absorptiometry and quantitative computed tomography for the mea- surement of bone mineral content. Palmer, S. B.; Langton, C. M. UI- trasonic studies of bone. IOP Publishing; 1987.

Hesp, R.; Klenerman, L.; Page, L. Decreased radial bone mass in Colles’ fracture. Acta Orthop. Stand. 573-557; 1984.

Horowitz, M.; Wishart, J. M.; Bochner, M.; Need, A. G.; Chatterton, B. E.; Nordin, B. E. C. Mineral density of bone in the forearm in pre- menopausal women with fractured wrists. Br. Med. J. 1314-1315; 1988.

Krolner, B.; Tondevold, E.; To& B.; Berthelsen, B.; Nielsen, S. P. Bone

mass of the axial and appendicular skeleton in women with Colles’ frac- ture: its relation to physical activity. Clin. Physiol. 147-157; 1982.

Langton, C. M.; Palmer, S. B.; Porter, R. W. Measurement of broadband ultrasound attenuation in cancellous bone. Eng. Med. 89-91; 1984.

Lauritzen, .I. B.; Schwartz, P ; McNair, P. Colles’ fracture and the risk of later hip fracture-a prospective follow-up study. Clin. Phys. Meas. 183; 1988.

Nilsson, B. E.; Westlin, N. E. The bone mineral content in the forearm of women with Colles’ fracture. Acta Orthop. &and. 836-844; 1974.

Owen, R. A.; Melton, L. J.; Ilstrup, D. M.; Johnson, K. A.; Riggs, B. L. Colles’ fracture and subsequent hip fracture risk. Clin. Orthop. Rel. Res. 37-43; 1982.

Petley, G. W.; Hames, T. K.; Cooper, C.; Langton, C. M.; Cawley, M. I. D. A comparison of single photon absortiometry and broadband ultrasound attenuation: past, present and future. Palmer, S. B.; Langton, C. M., eds. Utrasonic studies ofbone. IOP Publishing; 1987.

Poll, V.; Cooper, C.; Cawley, M. I. D. Broadband ultrasound attenuation in the OS calcis and single photon absortiometry in the distal forearm: a comparative study. Clin. Phys. Physiol. Meas. 7(4):375-379; 1986.

Salisbury, P. J.; Porter, R. W. Measurement of lumbar sagittal mobility, a comparison of methods. Spine 190- 193; 1987.

Wasnich, R. D.; Ross, P D.; Heilbrun, L. K.; Vogel, J. M. Prediction of postmenopausal fracture risk with use of bone mineral measurements. Amer. .I. Obs. Gym. 745-751; 1985.

Received: March 23, 1989 Revised: August 18, 1989 Accepted: December 7, 1989