Risk Factors for Osteoporosis: A Review

Janet R. Guthrie, MSc, Dip Ed, PhD,Lorraine Dennerstein, AO, MBBS, PhD, FRANZP, DPM and John D. Wark, MBBS, PhD, FRACP

In This Article

Risk Factors

Estrogen deficiency occurring after natural or surgically induced menopause leads to an uncoupling between osteoclasts and osteoblasts, which is responsible for accelerated loss of bone. This menopausal bone loss is predominantly trabecular[2] and also can occur in women before the menopause if estrogen levels fall before the final menstrual period.[3] It has been suggested that estrogen deficiency increases bone resorption partly by causing increased paracrine production of bone resorbing cytokines.[4] These cytokines appear to increase bone resorption by stimulating the development of osteoclast progenitors and increasing the activity of mature osteoclasts.[5,6,7,8,9]

Hormones other than estrogen have been reported to influence bone formation and resorption. Adequate androgen levels are most likely necessary for attainment of peak BMD and maximization of mechanical bone strength.[10] Although androgen levels do start to decline in women with increasing age,[11] there is no definite conclusion about the role of androgens in menopausal bone loss.

There are data supporting a proposal that progesterone has a stimulating effect on bone formation,[12,13] but this evidence comes from women with abnormal menstrual cycles who took progesterone. A study[12] of pre- and perimenopausal women found no association between BMD or rates of bone loss and progesterone levels.

The role of sex hormone binding globulin (SHBG) in the determination of BMD is uncertain.[14] SHBG has been found to be significantly higher in women with osteoporotic fractures compared with controls.[15] Levels of SHBG are increased by estrogens and decreased by androgens, and any effect of SHBG on BMD may be mediated through the levels of these hormones.

The majority of studies that have dealt with postmenopausal bone loss indicate that BMD depends on the number of years since menopause and not on chronologic age.[16,17,18,19,20,21] Postmenopausal women who have experienced an early menopause have significantly lower BMD than those whose age of menopause was normal. Nordin and colleagues[22] suggested that an early menopause is not a significant risk factor for osteoporosis, and Luisetto and associates[23] showed that after the age of 74 years, any difference in BMD in women experiencing an early or normal menopause has disappeared. There are limitations to this latter study: first, it is a cross-sectional study, and the only association that can be assumed is the one between the 2 variables, BMD and time; second, it was not a random sample; and third, BMD was measured at the distal forearm, where bone is mainly cortical, and there is no general agreement that cortical bone measurements are predictive of postmenopausal bone loss. Another consideration is that the numbers in the older age groups were smaller, and the sample may have been biased by deaths due to other causes in both groups -- that is, survival of the fittest might not have been related to their BMD but to some other factor in those women who had had an early menopause. The consensus of opinion favors an early menopause as being a risk factor for osteoporosis.

Various measures of body size have been shown to be associated with BMD, in particular body weight and height, which have been positively associated with BMD,[24,25,26] but there is controversy as to which features of body size are the most important for women in the pre-, peri- and postmenopause. Cross-sectional data have found that both total body fat[27] and lean tissue mass[28] explain a large percentage of the variation in BMD in postmenopausal women. Reid and colleagues,[29] in a 2-year prospective study of 122 normal postmenopausal women, reported that the rate of total body bone loss was significantly directly related to fat mass and the rate of change in fat mass.

The positive relationship between increased weight and increased BMD is probably the result of increased mechanical forces on the bone.[30] Because lean and fat mass are the 2 major components of body weight, it is not surprising that fat and lean mass have both been found to be positively related to BMD.[25,27,31,32,33,34] There is some discordance among these studies as to the relative importance of lean or fat mass in predicting bone mass. The differences in study results may reflect differences in study design; for example, in Aloia's study,[33] obese women were excluded. It is hypothesized that different mechanisms may predominate at different stages of life.[28] Salomone and associates,[35] in a cross-sectional study of pre- and perimenopausal women, found that body weight alone is not associated with increased bone mass unless a significant proportion of that weight consists of lean mass. Reid and colleagues[32] found a codependence of BMD on fat and lean mass in premenopausal women, but defining the relationship BMD/height as an index of "true" density, the association was with fat mass alone. A genetic association as well as physical activity may underlie the relationship between lean mass and BMD.[36,37,38,39] However, interpretation of studies of physical activity is difficult because such studies often include self-selected individuals who represent extremes in physical activity, and these results are difficult to extrapolate to the general population.[40,41,42,43]

The issue of collinearity is a problem when examining the influence of anthropometric variables on BMD, as areal BMD depends on body size. As mentioned above, Reid and colleagues[29] used BMD/height to compensate for differences in body size in premenopausal women, with a resulting change in associations. Also, Reid and colleagues[44] measured the volumetric BMD of the third lumbar vertebrae from simultaneous anteroposterior and lateral scans using dual-energy x-ray absorptiometry (DXA) and found an association between this measure of BMD and fat mass similar to that shown with the areal measures.

Co-twin difference analyses provide an important method of identifying the anthropometric determinants of BMD.[45] Because of the high correlation of body size between twins, the within-pair twin analysis reduced collinearity between lean and fat mass. Flicker and coworkers[34] found within-pair differences in lean mass to be independently associated with BMD at all sites in 69 postmenopausal sets of twins (aged 60-89 years). By contrast, fat mass was not independently associated with BMD at any site after allowance had been made for lean mass. This is in contrast to the findings of Reid and associates[32] of unrelated premenopausal women using BMD/height as a measure of bone density. Wark[46] reported that in a co-twin analysis of middle-aged pre- and early postmenopausal twin pairs, fat and lean mass were independently related to total body bone mineral content. BMD of the lumbar spine, proximal femur, and femoral neck was associated with lean mass in twins concordant for premenopausal status and with fat mass in twins concordant for postmenopausal status.

It does appear that the association between fat and lean mass and BMD probably depends on the age and menopausal status of the population studied. Conversion of androgens to estrogens becomes a major source of estrogen only after the menopause.[47] The endocrine effect of fat mass on bone may be minimal before menopause. After the menopause, fat mass and weight increase while lean mass decreases,[48] and the hormonal contribution of the fat mass may explain the association between fat mass and BMD.[49] Adipose tissue contains aromatase, and this enzyme is responsible for the conversion of androgenic steroids into estrogens -- the greater the fat mass the greater the amount of estrogen synthesized. The association of BMD with lean mass in elderly twins[34] supports the need for further study of the relationship between body composition and bone strength in the elderly.

Physical activity - The effect of physical activity on bone mass has been studied extensively. The first evidence of an association came from studies of the effect of inactivity on bones, particularly the osteopenic effects of weightlessness in astronauts during space flights.[50] Physical activity has also been recognized as influencing the attainment of peak bone mass in childhood and adolescence.[51,52,53] The role of physical activity in maintaining BMD during the menopausal transition and in the postmenopausal years is still a subject of contention. There are several reviews on the effects of exercise on BMD in women,[54,55,56,57,58,59,60] and the results are not consistent.

Cross-sectional studies on the effect of physical activity on BMD are subject to inherent bias. A recent study[61] reported that midlife women who were habitually exercising vigorously more than twice a week had higher total body bone mineral content than those women involved in light to moderate activity no more than once a week. These currently physically active women in Uusi-Rasi's[61] study had also been physically active throughout their lives, so it is difficult to evaluate in this cross-sectional study whether the observed bone mineral content resulted from benefits accrued from exercise during growth and development or continued habitual physical activity or, as is most likely, both. There are very little data with regard to longitudinal (cohort) studies. The best evidence that exercise can slow bone loss or add bone mass to the postmenopausal skeleton comes from prospective intervention studies -- although the amounts are site specific and relatively modest, and the activity must consist of high-load resistance exercise.[62,63,64] This evidence, however, comes from trials in which small numbers were involved, and the studies were short in duration and thus have the potential for selection and other biases.

Calcium - The data relating to the benefit of high calcium intake during the menopause transition and after the menopause are inconclusive. The results of studies concerning the relationship between dietary calcium intake and postmenopausal bone loss and the effect of calcium supplementation on postmenopausal bone loss have not been concordant. Heaney[65] reviewed 43 studies published since 1988 that related calcium intake to bone mass, bone loss, or bone fragility. Twenty-six reported that calcium intake was associated in some way with bone mass, bone loss, or fracture, and 16 did not. The controlled intervention studies suggest that calcium supplementation (1000 mg or more) can slow down postmenopausal bone loss. This effect can be more easily demonstrated in women who are more than 5 years postmenopausal.[65,66] Calcium supplements have been observed to be most effective in those women in whom the baseline calcium intake was low, the mean age was high, and there was clinical evidence of osteoporosis.[67,68]

In summary, the bulk of published research data indicates that calcium has a beneficial effect at some bone sites in controlled trials where high doses of calcium ensure that one group is getting much more than a control group. The effect of added calcium on femoral neck BMD in postmenopausal women is not yet established. Second, the data seem not to support a positive effect of calcium on bone loss in women in the peri- and early postmenopausal years.[65,69,70,71] The maximal effect of a calcium supplement appears to occur with a dosage of approximately 1000 mg per day,[70] although no published study is able to say how much calcium is actually needed.

An active metabolite of vitamin D [1,25-dihydroxyvitamin D, 1,25(OH)2D3] is necessary for the normal absorption of calcium from the intestine, and hence a deficiency of this vitamin is a risk factor for osteoporosis. This compound can be formed in the skin, and the rate of synthesis is increased when the skin is exposed to sunlight. Low vitamin D status is attributed to low dietary intake, decreased skin exposure to the sun, or impaired synthesis of vitamin D in the skin.[72] Vitamin D deficiency is a problem in institutionalized elderly people, in people not exposed to sufficient sunlight,[73] and particularly in countries where fortification of food with this vitamin is not practiced.[74]

Estrogen and 1,25(OH)2D3 appear to have common activities in the intestine. It has been shown in animal studies[75] that estrogen stimulates the expression of the 1,25(OH)2D3 receptor that drives the calbindin-mediated transport of calcium across the intestinal epithelial cells. It is possible that the negative effect of vitamin D deficiency on the bones of aged patients is enhanced by this lack of estrogen-dependent expression of the 1,25(OH)2D3 receptor.

Caffeine, smoking, alcohol - Dietary caffeine induces a negative calcium balance through increased urinary loss,[76] but information about the influence of caffeine intake on bone mass is limited. No association was evident in some studies,[77,78,79] but an adverse effect was observed in others.[80,81]

Smoking has also been reported in both cross-sectional and longitudinal studies to be a risk factor for vertebral, forearm, and hip fractures and to be associated with lower BMD.[33,82,83,84,85,86] The adverse effect of smoking on bone is likely to be mediated through changes in endogenous estrogen metabolism; estrogen production is decreased and metabolic clearance has been reported to be increased in smokers,[1] but this has not been confirmed in all studies.[87] Daniell[82] hypothesized that lower body weight of smokers and direct or indirect effects on bone resorption are responsible for the detrimental effect of cigarettes on bone mass. Aloia and coworkers[84] reported a negative association between cigarette smoking and calcium absorption. Smoking is also known to produce an earlier menopause,[88] and an early menopause is a risk factor for osteoporosis.

Data pertaining to the effects of moderate amounts of alcohol on bone metabolism and osteoporosis are limited; evidence from population studies has been conflicting, with most finding no effect of alcohol intake on BMD in pre- and postmenopausal women.[27,89,90,91,92,93,94,95,96] Several studies[97,98,99,100] have reported a positive association between moderate alcohol intake and BMD in postmenopausal women. In an analysis of women aged 68-96 years in the Framingham Study, Felson and associates[99] found that after adjusting for age, BMI, and smoking, women who drank at least 16 standard alcoholic drinks per week had on average 5% to 10% higher BMD at the radius, spine, and proximal femur than women who drank less than 2 standard drinks per week. Baudoin and colleagues[100] found that moderate alcohol use (11-29 g/day) was associated with a significant increase in trochanteric BMD in women aged 75 years and older. This association between alcohol intake and BMD may result from increased conversion of androstenedione to estrone, but more hormonal studies in postmenopausal women are needed to confirm this finding.

Gynecologic variables, such as oral contraceptive (OCP) use, parity, breast feeding, age of menarche, and menstrual cycle irregularities, have all been associated with BMD. OCP use has been associated with an increase or no effect on BMD, so it is not a risk factor for osteoporosis and will not be discussed further. Studies of the relationship between parity and BMD have been conflicting: Parity has been determined to have a positive association,[84,90,101,102,103] a negative association,[104,105] and no[16,106] association with BMD.

Similar discrepancies have been found in relation to breast feeding and BMD. Some studies[84,102,107] have reported a positive association between BMD and breast feeding and others a negative association[101,105,108,109,110] or no association.[106,111] Dequecker and coworkers[107] reported that BMD depended on the length of lactation: women who had a total lactation time of longer than 6 months showed a significantly lower bone mass at the trabecular area of the distal radius compared with those who had a lactation time of less than 6 months.

The studies mentioned above have not always adjusted for potentially confounding covariates such as age, time since menopause, obesity, hormone therapy use, and cigarette smoking. Another problem is whether allowance was made for use of calcium supplements or high dietary calcium intakes during pregnancy and lactation. Kritz-Silverstein and associates[112] examined the independent association of number of pregnancies and breast feeding in a large community-based sample of older postmenopausal women who would have completed their pregnancies and menopause before there was widespread use of calcium supplements. They reported that unadjusted BMD values at the wrist, radius, and hip increased with increasing number of pregnancies, and women who had breast fed had higher BMD at these sites. However, after adjustment for age, age at menopause, obesity, cigarette smoking, and estrogen and thiazide use, these associations were no longer significant. They concluded that reproductive history is not a long-term determinant of BMD.

Women with late onset of menarche have been reported to have significantly reduced peak bone mass[113] and increased fracture risk.[114] Ito and colleagues[115] showed that BMD of the lumbar spine in postmenopausal women was strongly associated with the length of reproductive years, so that women who had an early menarche and late menopause were advantaged. Women with high BMD and early menarche usually have higher body weight, so it is difficult to separate the effects of anthropometric factors and age of menarche on BMD.

Menstrual cycle irregularities may also have an effect on BMD. The effects on bone of oligomenorrhea and periods of amenorrhea have been documented mainly in cross-sectional studies,[116,117,118,119] which reported lower values for BMD than in control populations. Relatively little is known about the relationship between menstrual cycle characteristics (cycle length, variability, and bleeding duration) and BMD in the general population of pre- and postmenopausal women. Long cycles may reflect relatively low estrogen levels, as they represent a larger proportion of time spent in the early part of the follicular phase during which estrogen and progesterone levels are low.[120] One prospective study[114] found that there was an increased fracture rate at the wrist, spine, and hip in women who had reported a cycle length of greater than 30.5 days at age 28 to 32 years. Highly variable cycle length (very long or very short cycles) may indicate anovulatory cycles, and this can reflect either hypoestrogenic or hyperestrogenic conditions.[121]

Peak bone mass of premenopausal women is a major determinant of subsequent risk of osteoporotic fracture. So maximizing bone mass during skeletal growth and development and maintaining it during the premenopausal years are important strategies in the prevention of osteoporosis. Low peak BMD and a high rate of bone loss have been shown to significantly predispose to fractures to the same extent, with an odds ratio of about 2.[122] If both of these conditions were present, the odds ratio was about 3.[121] Kroger and colleagues[123] reported that in a population sample of perimenopausal women, those whose spinal BMD was in the lowest quartile had a 2.9 times greater risk of future fracture than those in the highest quartile. The World Health Organization has proposed diagnostic categories for thresholds of BMD on the basis of the distribution of skeletal mass in young healthy individuals. If BMD is more than 1 standard deviation below the young adult mean, treatment should be considered to prevent bone loss. If it is more than 2.5 standard deviations below, treatment is strongly advised.

One of the major determinants of peak BMD is the genetic components that contribute to bone formation and resorption. The magnitude of the effect of the genetic component appears to vary according to age and site. Twin studies estimate that 60% to 80% of the variance in BMD is due to a genetic component.[39] In their studies of families, Krall and Dawson-Hughes[124] found that only 46% to 62% of the variance in BMD could be attributed to heredity. Lifestyle and hormonal factors discussed above as risk factors for osteoporosis also have an effect on the attainment of peak bone mass.

One genetic component of bone mass has been ascribed to allelic variation in the receptor for 1,25 dihydroxy-vitamin D, one of the controllers of calcium metabolism. These vitamin D receptor (VDR) alleles have been reported to have a codominant effect on bone turnover[125] and a codominant effect on BMD.[126] The results of Morrison's studies have not been confirmed by all other studies to date, and it is uncertain whether the VDR genotype influences peak bone mass alone or the rates of bone loss that occur with aging. Cooper and Umbach[127] used a meta-analytic approach to assess the association between VDR and BMD. Twenty-nine papers, published up to the end of July 1996, were assessed with regard to this association and the influence of specific study characteristics, such as skeletal site measured, age, and menopausal status of subjects. The authors compared homozygous genotypes and reported a greater effect size (mean difference in BMD divided by standard deviation) at the hip than at the spine but concluded that "VDR polymorphisms represent one genetic factor affecting BMD, but further research into the mechanisms, clinical significance, and its relation between other genetic and environmental factors is needed."

With respect to race, black women in the United States have been found to have greater BMD than white women,[128] and black women have also been reported to experience fewer osteoporotic fractures than white women.[129,130] Differences in hip fracture rates have been reported between various European populations who are from the same racial origins.[130] The reasons for these geographic differences in the incidence of hip fracture are not fully understood,[131] but they are possibly related to environmental or a combination of genetic and environmental factors.

In conclusion, as women age and experience the menopausal transition, the risk of developing osteoporosis increases. In addition, low body mass index, low calcium intake, low level of physical activity, and smoking can affect BMD. The relative importance of these factors on BMD in midlife women is not fully established. The impact of gynecologic history (parity, lactation, oral contraceptive use, age of menarche) on BMD is uncertain. Many factors have been implicated as influencing bone loss, but the evidence for some is unsubstantial.

Some of the risk factors for osteoporosis are nonmodifiable, such as genotype, age, and race. Others, such as calcium intake, smoking, and physical activity, are potentially modifiable. In Australia, advertising campaigns by the Australian Dairy Corporation and Osteoporosis Foundation target the lifestyle habits of middle-aged women. The changes, beneficial or otherwise, that do occur in these modifiable risk factors as women pass through the menopausal transition have yet to be investigated.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: