Premenopausal bone loss in the lumbar spine and neck of femur: A study of 225 Caucasian women

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

Bone, 11, l-5 (1990) Printed in the USA. All rights reserved.

Premenopausal Bone Loss in the Lumbar Spine and Neck of Femur: A Study of 225 Caucasian Women A. RODIN,’ B. MURBY,2 M. A. SMITH,3 M. CALEFFL4 I. FOGELMAN2 1Departments of Gynaecology, 2 Nuclear Medicine, 3 Radiological Sciences and 4 Oncology, Guy’s Hospital, London, Address for correspondence

I. FENTIMAN,4

M. G. CHAPMAN’

and

England

and reprints: Dr. Adam Rodin, Department

of Gynaecology, Guy’s Hospital, London, SEI, England.

Abstract Two hundred and twenty-five premenopausal women were studied to evaluate age-related changes in trabecular bone mass. Measurements were made at the lumbar spine and femoral neck by dual photon absorptiometry. It was found that spinal bone density increased signifcantly from the 20s to reach a peak in the mid30s. Identical trends were observed in total bone mass and bone mass normalized by length. Bone loss then proceeded at a rate of 1% per year, and by the early SOS, 10% of peak spinal density was lost. There was no peak in femoral neck density; loss commenced in the late 20s and continued at a rate of 0.4% per year. The cumulative premenopausal loss from the femur at 9% was comparable to that in the spine. It is concluded that signIficant amounts of trabecular bone are lost from both the spine and femoral neck before the menopause. The implications of these findings for the prevention of osteoporosis are discussed. Key Words: Bone density-Dual Menopause.

Table I. Characteristics

Introduction Age is well recognized as an important determinant of bone density, but the natural history of age-related bone loss remains controversial. In women, cortical bone density shows little or no diminution until the menopause which is followed by a period of rapid bone loss (Riggs et al. 1981). Advances in technology have led to the introduction of dual photon absorptiometry and quantitative CT scanning and these techniques make it possible to measure bone density at the spine and femur, the clinically important sites where osteoporotic fractures tend to occur. These areas have a significant trabecular bone component and evidence is accumulating that cortical and trabecular bone exhibit different patterns of loss. Three patterns of age-related changes in trabecular bone have been described. Riggs et al. (1981) demonstrated loss commencing in “young adulthood” and continuing in a linear manner throughout life. A second model suggests that trabecular bone behaves in a similar way to cortical bone; it is maintained until the menopause when loss commences (Aloia et al. 1985; Sambrook et al. 1987). Other groups have shown that trabecular bone mineral reaches a peak in the mid-30s and this is followed by a progressive

photon absorptiometry-

of the study population.

Age group (years)

n

Height (m) mean 2 SD

18-22 21-25 24-28 27-31 30-34 33-37 36-40 39-43 42-46 45-49 48-52

55 69 41 33 27 19 28 30 23 17 15

1.65 1.63 1.65 1.63 1.66 1.62 1.62 1.62 1.63 1.60 1.56

? 2 ? ? * 2 ” 2 ? ” ”

Weight (kg) mean f SD 59.30 58.90 57.58 61.79 63.44 62.17 57.00 62.20 64.83 65.90 75.00

0.06 0.14 0.06 0.21 0.06 0.08 0.05 0.05 0.05 0.04 0.02 1

f * + 2 f + 2 ? 2 ? 2

8.07 7.11 12.0 10.0 9.59 11.5 17.0 9.99 8.36 11.2 14.3

Pill use (year) mean 1.5 2.4 2.9 2.9 3.1 2.3 2.0 1.8 2.0 2.5 1.9

2

loss (Krolner and Pors-Nielsen 1982; Cann et al. 1985). The aim of the present study is to establish the pattern of agerelated alterations in trabecular bone in normal Caucasian women. An understanding of these changes may suggest alternative approaches to the prevention of osteoporosis.

A. Rodin et al.: Premenopausal

bone loss in spine and femur

pendent of normalization, non-linear regression analysis was also performed on the BM and BM/L data in the spine. The significance of the difference between BMD at different ages was assessed using Students t test for unpaired data. To further analyze the significance of any visually identifiable peak linear regression, analysis was performed and rate of change in bone density was calculated.

Patients and Methods Results Patients The study population consisted of 225 Caucasian women whose ages ranged from 18 to 52 years. Women were recruited from the hospital staff and from outpatient clinics; approximately 50% of volunteers in each age group were staff members. They were all premenopausal with regular menstrual cycles. Subjects had no relevant medical history, and in particular, there was no history of bone or joint disease. None of the women were taking medication known to influence bone density. Women with a daily calcium intake of less than 500 mg, as assessed by questionnaire, were excluded as were those with a history of anorexia nervosa. In all cases, body mass index was ~25 (weight in kilograms/(height in meters)*). The characteristics of the study population are shown in Table I. Height and weight were similar in all age groups with the exception of the oldest group which was significantly shorter and heavier than other age groups. Current and past pill users were included in this study; we have previously shown that combined oral contraception does not influence bone density (Rodin et al. 1987). All subjects gave informed consent for the study and approval was granted by the Ethical Committee, Guy’s Hospital, London. Methods

Second and higher order polynomial fits confirmed the existence and position of a peak in spinal BMD. Further confirmation was provided by the data on BM and BM/L (Fig. 1). The data were significantly correlated with a curve with a peak in the mid-30s (r = 0.24; p < 0.001). There was no

‘O< l

60

.

t

.

l*.*

.

I**

:-

-- . 10

20

30

5'0

4b

t 0

AGE

-

I

.

?6-

l.

.

.

. .

l

. .

S-

Individuals attended for densitometry on entry to the study and there was no grouping of measurements with time according to age. All scans were performed and analyzed by the same technician. Bone mineral was measured at the lumbar spine (L2-L4) and at the neck of femur by dual photon absorptiometry using a Novo 22a dual photon absorptiometer. Total bone mineral (BM), bone mineral per unit length (BM/L) and bone mineral per unit area or bone mineral density (BMD) were measured at both sites. The units used were respectively, grams hydroxyapatite (gHA), grams hydroxyapatite per unit distance between the top of L2 and the bottom of L4 (gHA/cm) and grams hydroxyapatite per unit area of L2 to L4 (gHA/cm2) as measured by the edge detection algorithm. Reproducibility data was obtained by repeated measurements in 10 subjects at both sites at 1Zweek intervals, giving a precision of 2.2% for BM, 2.6% for BM/L and 2.0% for BMD in the lumbar spine, and 2.2% for BMD in the femoral neck. In the analysis of the data subjects were grouped according to age. Overlapping 5 year groups were formed (18-22 years; 21-25 years; 24-28; 27-31; 30-34; 33-37; 36-40; 39-43; 42-46; 45-49 and 48-52). The mean BMD and standard error of mean were calculated for each group and plotted against the mean age for that group. Non-linear regression analysis was also performed on the whole data set. In order to confirm that the observed trends were inde-

2! 10

1

I

8

I

20

30

40

50

I 40

I 50

60

AGE

:::

0.6 f 10

I 20

1 30

. 60

AQE

Fig. 1. Spinal BM, BM/L and BMD plotted against age for 225 premenopausal women. Second order polynomial fit shown.

A. Rodin et al.: Premenopausal bone loss in spine and femur

3

correlation with a linear fit. There was a 5% increase in spinal bone density between the age groups 18-22 and 33-37 which was significant (p = 0.03). This was followed by a progressive decline at a rate of 1.O% per year. By the early 50s 10% of peak spinal bone mass was lost and this represented a significant fall (I, = 0.016) (Fig. 2). A different pattern was exhibited by the femoral neck, the raw data of which showed a linear correlation with age (r = 0.31, p < 0.001). No improvement in fit was obtained using a second or higher order polynomial. Our study did not demonstrate a peak in femoral bone mass and loss of mineral commenced in the late 20s. The rate of loss was slower than that seen in the lumbar spine at 0.4% per year, however, by the early 5Os, 9% of femoral bone mass was lost which represented a significant premenopausal fall @ = 0.012) (Fig. 3).

and is linear. However, the bulk of evidence now conflicts with this early model. Krolner and Pors-Nielsen (1982) examined lumbar spine bone mineral content in 70 women in relation to age. By fitting the data to a gamma variate function they showed a maximum spinal density at 34 years, but this method of statistical handling has been criticized (Tothill et al. 1983). Their own results support the view that there is increased loss of spinal bone mineral after the menopause, but insufficient numbers prevent detailed analysis in premenopausal women. This limitation also applies to other studies. Nilas and Christiansen (1987) carried out a cross-sectional study of 178 healthy women and found no evidence that substantial premenopausal bone loss occurs from any site. Hansson and Roos (1986) in agreement with our findings, noted a continuous age-related decrease of spinal bone density after the age of 35. Mazess et al. (1987) with a heterogeneous study population of 892 women derived from seven centers found a 10% decrease in spinal density in the decade preceding the menopause. A similar pattern was shown for femoral bone density. Schaadt and Bohr (1988), in a recent study, found that the bone mineral content of the femoral neck decreased linearly from the 30s. Premenopausal bone loss is also suggested by histomorphometric studies (Meunier et al. 1973; Marcus et al. 1983) and by postmortem investigations (Weaver and Chalmers 1966; Arnold 1973). The differing descriptions of age-related changes in trabecular bone are probably explained by technological differences, differences in data handling and differences in the sample sizes between the groups. Sambrook et al. (1987) pointed out the pitfalls of using cross-sectional data to draw conclusions about longitudinal changes in a population, and they emphasized the importance of adequate sample size. Our data and results from previous studies provide a compelling body of evidence that significant loss of trabec-

Discussion In this study, we have demonstrated that significant premenopausal bone loss occurs from both the lumbar spine and neck of femur. Spinal bone density increases in the 20s to reach a peak in the mid-30s and then falls in an almost linear fashion. The femoral neck shows a different pattern of bone loss; peak density is presumably attained at an earlier age which cannot be defined from the present data. The rate of loss is slower than that seen in the spine, but as bone loss commences at an earlier age, by the late 40s the cumulative loss from peak density at both sites is comparable . Previous studies offer conflicting evidence about the occurrence of premenopausal bone loss and the patterns of age-related changes in bone mineral. In a cross-sectional study of 105 normal women, Riggs et al. (1981), found that bone loss from the vertebrae begins in young adulthood

SPINAL BONE DENSITY

v AGE IN PREMENOPAUSAL

WOMEN

Bone density 6 gHA/cm* 1.00

Year running average

plotted +- SEM

_ n =225

0.96

_

0.92

_

0.88 _

I

I

T

1

I

0.84 _

o.ao 1

~--p=O.O33~&p=O.O16-~

I 20

I

I 24

I

I,

I

28

32

I,

IT 36

I 40

I 44

I

I 48

I

I 52 Age

BMD plotted against age in plotted against the mean age for that group.

Fig. 2. Spinal

225

premenopausal

women.

Mean

BMD

was calculated

bears)

for 5 year overlapping groups and

A. Rodin et al.: Premenopausal Ban3 density

bone loss in spine and femur

FEMORAL BONE DENSITY v AGE w PREMENOPAUBAL WOMEN

gtiA/cm2 5 year running average pbtted + SEM

I

0.88 _

0.84 _

11

I

t-t=225

111

0.80 _

II 0.78_

0.72_

0.88 _

0.84 _

k-,=0.012.-4

,

I

20

24

28

I1

1

32

I1

I.

11

I1

81

38

40

44

48

52 Age (years)

Fig. 3. Femoral BMD plotted against age for 225 premenopausal plotted against the mean age for that group.

ular bone occurs before the menopause. The finding of a peak in spinal bone density is supported by similar trends in BM and BM/L suggesting that this phenomenon is independent of area normalization. The significant increase in spinal bone density between the 20s and mid-30s may have important implications for the prevention of osteoporosis in later life. In addition, our study provides further evidence that trabecular bone at different anatomical sites exhibits different patterns of age-related density changes, but the reason for this variation is not fully understood. The possibility that these patterns of changes in trabecular bone are biased by secular trends needs consideration. It seems unlikely, however, that the rise in spinal density seen between 18 years to the mid-30s could be due to such influences. The relatively short time scale argues against a secular effect. However, we cannot rule out the possibility that the differences between those at the age extremes of the study population are influenced by secular effects. The mechanism of early bone loss remains unclear. The importance of oestrogen in maintaining skeletal bone mass is well established and there is evidence for accelerated bone loss after oophorectomy (Cann et al. 1980) or following a natural menopause (Gennant et al. 1984). Oestrogen replacement therapy is effective in preventing postmenopausal bone loss (Lindsay et al. 1976). However, the menopause is preceded by a transition phase which is characterized by rising gonadotrophin levels and falling total oestrogen output from the ovaries (Sherman et al. 1976). It therefore seems likely that the patterns of endocrine changes will vary between individuals, and we suggest that genetic and environmental factors interact to influence the timing and profile of the hormonal fluxes associated with the menopause. A proportion of women will therefore becomge relatively oestrogen deficient before the cessation of their menses, and it is well known that some develop

women. Mean BMD was calculated for 5 year overlapping groups and

vasomotor instability and genital atrophy during the climacteric. It has been suggested that this relative deficiency of oestrogen may contribute to the loss of trabecular bone in premenopausal women (Johnston et al. 1985). It is possible that this effect contributes to spinal bone loss which commences in the mid-30s but it does not adequately explain the pattern of loss seen in the femoral neck. Prevention of osteoporosis is an important goal, and while hormone replacement therapy in the early premenopausal years is of proven efficacy (Lindsay et al. 1976), our findings raise the possibility that prophylaxis in premenopausal women should be considered. Two potential approaches are suggested: (a) spinal bone density increases in the young woman until the mid-30s and if this can be enhanced the peak spinal bone density can be maximized; (b) if bone loss in the climacteric can be prevented or reduced, women would again reach the menopause with a better reserve. Exercise and calcium supplementation may be suitable strategies, but before any measures can be recommended they must be fully evaluated and ultimately they must be shown to reduce fracture risk.

Acknowledgment:

We would like to acknowledge the assistance of Bernice Threadgold who performed the densitometry.

References Aloia, J. E.; Vaswani, A.; Ellis, K., et al. A

model for involutional loss. 1. Lab. C/in. Med. 106:630-637; 1985.

bone

Arnold, J. S. Amount and quality of trabecular bone in osteoporotic vertebral fractures. Clin. Endocrinol. Metab. 2:221-238; 1973.

and

Cann,

C. E.; Gennant,

H. K.; Ettinger,

B.; Gordon,

G. S. Spinal

mineral

A. Rodin et al.: Premenopausal loss in oophorectomized

bone loss in spine and femur

women. JAMA 244:2056-2059;

1980.

Cann, C. E.; Genant, H. K.; Kolb, F. 0.; Ettinger, B. Quantitative computed tomography for prediction of vertebral fracture risk. Bone 6: I-7; 1985. Gennant, H. K.; Cann, C. E.; Ettinger, B., et al. Quantitative computed tomography for spinal mineral assessment. Christiansen, C.; Amaud, C. D.; Nordin, B. E. C., et al., eds. Osreoporosis I. Aalborg Stiftsbogtrykkeri; 1984:65-71. Hansson, T.; Roos, B. Age changes in the bone mineral of the lumbar spine in normal women. Calcif. Tissue Jnr. 38249-251; 1986. Johnston, C. C.; Hui, S. L.; Witt, R. M., et al. Early menopausal changes in bone mass and sex steroids. J. Clin. Endocrinol. Merab. 61:905-911; 1985.

Krolner, B.; Pors-Nielsen, S. Bone mineral content of the lumbar spine in normal and osteoporotic women. Clin. Science 62329-336; 1982. Lindsay, R.; Hart, D. M.; A&ken, J. M., et al. Long-term prevention of postmenopausal osteoporosis by oestrogen. Lancer 1:1038-1041; 1976. Marcus, R.; Kosek, R. J.; Pfefferbaum, A.; Homing, A. Age-related loss of trabecular bone in premenopausal women: a biopsy study. Calcif. Tissue Inr. 35:406-409, 1983. Mazess, R. B.; Barden, H. S.; Ettinger, M., et al. Spine and femur density using dual-photon absorptiometry in US white women. Bone and Mineral 2:211-219; 1987. Meunier, P.; Courpron, P.; Edouard, C., et al. Physiological senile involution and pathological rarefaction of bone. C&z. Endocrinol. Merab. 2~239; 1973.

Nilas, L.; Christiansen, C. Bone mass and its relationship to age and the menopause. J. Clin. Endocrinol. Merab. 65~697-702; 1987.

5

Riggs, B. L.; Wahner, H. W.; Dunn, W. L., et al. Differential changes in bone mineral density of the appendicular and axial skeleton with aging. J. Cfin. Invesr. 67~328-335; 1981. Rodin, A.; Fogelman, I.; Chapman, M. G. Combined oral contraception as a determinant of bone mass. Abstracts from the Inrernarional Symposium on Osteoporosis, Denmark. Osteopress; 1987:54. Sambrook, P. N.; Eisman, J. A.; Furler, S. M.; Pocock, N. A. Computer modeling and analysis of cross-sectional bone density studies with respect to age and the menopause. J. Bone Min. Res. 2: 109- 114; 1987. Schaadt, 0.; Bohr, H. Different trends of age-related diminution of bone mineral content in the lumbar spine, femoral neck and femoral shaft in women. Cal& Tissue Znr. 42:71-76; 1988. Sherman, B. M.; West, J. H.; Korenman, S. G. The menopausal transition; analysis of LH, FSH, Estradiol and progesterone concentrations during menstrual cycles of older women. J. Clin. Endocrinol. Merab. 42:629636; 1976.

Tothill, P.; Smith, M. A.; Sutton, D. Dual photon absorptiometry of the spine with a low activity source of gadolinium 153. Br. J. Radiology 56:829-835; 1983. Weaver, J. K.; Chalmers, 1. Cancellous bone; its strength and changes with aging and an evaluation of some methods for measuring its mineral content. ‘I. Age changes in cancellous bone. J. Bow Joinr Surg. 48A:289; 1966.

Received: July 21, 1988 Revised: February 2, 1989 Accepred: February 8, 1989