Reproductive Hormones and Skeletal Health in Young Women

In young women, low estrogen levels complicate a wide variety of diseases, including premature ovarian failure, anorexia nervosa, athletic amenorrhea, prolactinoma, hypopituitarism, and chronic kidney disease. Hypoestrogenemia may also result from therapy with GnRH agonists, glucocorticosteroids, chemotherapy, and especially aromatase inhibitors. All of these situations are associated with bone loss, because estrogen chronically suppresses bone resorption. Several mechanisms are involved: 1) estrogen increases osteoclast apoptosis; 2) by suppressing interleukins and proinflammatory cytokine expression in bone marrow cells, estrogen decreases the number of osteoclasts; 3) by inhibiting the production of receptor activator of nuclear factor-κB ligand, estrogen reduces the number and activity of osteoclasts; and 4) by increasing stromal cell/osteoblast cell expression of TGFβ, estrogen inhibits osteoclast activity. Estrogen also has some positive effects on bone formation by acting as a mitogen to cells early in the osteoblast line, reducing apoptosis of osteoblasts, and increasing expression of TGFβ, bone morphogenetic proteins, and IGF-I (1).

In these pathological conditions, the cause of the bone loss is frequently multifactorial. For example, women with anorexia nervosa also have low IGF levels. Nonspecific factors such as poor diet, reduced exercise, or chronic inflammation may further increase bone resorption.

Low estrogen levels are also seen in healthy young women who are using depo-medroxyprogesterone acetate (DMPA) for contraception. Bone loss is a side effect of this medication (2). The bone mineral density (BMD) decreases at a more rapid rate during the first years of therapy and then continues to decrease at a slower rate (3). After 5 yr, the loss is about 6% (0.5 sd). The DMPA effects are worse in teenagers; combined results from four prospective studies found an average loss of spine BMD of 3.1% over 2 yr in DMPA users compared with a gain of 7.2% in untreated controls (4). A cross-sectional study in this issue by Walsh and colleagues (5) focused on the age of starting DMPA. One group aged 18–25 started DMPA before age 20, and another group aged 35–45 started after age 34. Each group had matched controls. Importantly, the average use was 37 months in both groups. This avoids the confounding between age and duration of use, which is encountered in other studies. The BMD was significantly lower than controls when DMPA was started in teenagers but not when it was started after age 34. Estrogen levels were lower in the DMPA users than the non-hormone users and were also lower in the younger women than in the older women. The authors concluded that estrogen levels mediated the effect of DMPA on bone markers because adding DMPA to a regression model reduced the correlation between markers and estrogen. These correlations were not very strong (r = −0.17 between N-telopeptide and estradiol-treated), so this is a tenuous conclusion, but it is bolstered by the findings from two other studies that adding estrogen to DMPA abolishes the negative effect on the skeleton (6,7). The age range of the older group of women is underrepresented in most of the previous studies of skeletal effects of DMPA.

These results lead to several questions about basic bone biology. How are adolescent bones different from mature bones? How does DMPA use compare to the natural state of low estrogen seen in healthy lactating women? Is bone completely restored in these situations? There are also important clinical questions about potential risk of fractures and the skeletal risks from other forms of contraception.

Bone in Adolescents

It is difficult to precisely define the age of skeletal maturity, because it occurs in different bones at different ages, and cancellous bone loss has a different temporal pattern than cortical bone. Measurements of the skeletal changes with growth depend on the technique used; dual-energy x-ray absorptiometry (DXA) cannot separate cortical from cancellous bone. The areal density measured by DXA increases with growth even when there is no change in volumetric density as measured by quantitated computed tomography. In the vertebral bodies, the quantitated computed tomography reaches the highest levels shortly after cessation of growth, whereas DXA measurements continue to show increases (8). By the mid-20s the decline in cancellous bone has begun. Cortical bone, however, is more stable until menopause (9).

Even if actual BMD is no longer increasing in late teenage years, the markers of bone formation and resorption are still higher than in older women and are higher yet with DMPA use (5,10). Bone loss is not predicted by high turnover but by the bone balance, which is the difference between bone formation and bone resorption rates. The available markers are not sensitive enough to calculate the balance; this is even more complicated during early adolescence, because modeling requires high formation and resorption, and it is not possible to determine whether serum markers came from modeling or remodeling. The markers decline with age in late adolescence but don’t reach a plateau until a decade after achieving maximum height, possibly because it takes that long for some of the growth plates to fuse.

Newly formed bone is different from older bone because it is not yet fully mineralized. The osteoblasts first form osteoid matrix, which has no mineral; then, after several weeks, mineralization occurs, rapidly at first and then more slowly until reaching mature mineralization at least 3 yr later. Adolescents will have more newly formed bone than adults, so the BMD will be lower even if the volume of bone is the same. This could explain some increases in BMD measured radiographically in the years after skeletal growth.

Adolescent bone physiology is also different from adult bone due to higher levels of growth hormones and IGF. The role of androgens and the possible anabolic effects of estrogens in this age range are still not well understood.

Bone Loss and Gain Associated with Pregnancy and Lactation

In healthy young women who become pregnant, the gonadal hormones and the calcium-regulating hormones undergo complex changes with resulting changes in the bone physiology (11). During early pregnancy, estrogen levels increase, bone formation and resorption decrease, and BMD may increase. Later, as the fetus requires more calcium, the 1,25-OH₂ vitamin D increases, as does bone resorption and bone formation, and the BMD is stable or may show slight loss. Lactation results in a rapid loss of BMD, irrespective of calcium intake. In some mammals, this loss of BMD is profound, averaging 50% of the maternal skeleton. In healthy women, the loss is approximately 1% per month for the first 6 months. The estrogen levels are low, and the high bone resorption is also enhanced by PTHrP, which is expressed in the breast in response to prolactin. Calcitonin may play a role in attenuating the loss (12). This bone loss is reversed after weaning. Estrogen increases, PTHrP decreases, bone resorption ceases, and the osteoblasts that were primed by the PTHrP and previous resorption are poised to fill in the resorbed cavities with new bone. The BMD increases rapidly after weaning and then slowly until it returns to the prepregnancy baseline by 18 months. Epidemiological studies of postmenopausal women have not found that parity is a risk factor for osteoporosis (13). Thus, the bone architecture probably is restored after weaning, although this has not been definitely verified.

Very little attention has been paid to these secondary peaks in bone mass that occur after lactation, but these could be important intervals because bone mass is naturally increasing and lifestyle factors could potentially have a beneficial impact.

When adolescents become pregnant, the sequence of bone gains and losses are superimposed on the development of the skeleton. It is not clear whether pregnancy and lactation in teenagers harm the skeleton. Some cross-sectional studies have found lower bone density in those who had been pregnant, but social-economic factors could have played a role. In young primates from a controlled breeding colony, the 6-month postweaning BMD of the females that had undergone pregnancy and lactation were no different from those who were housed in cages without males (14).

The rapid increase in BMD after discontinuation of DMPA resembles the postweaning interval (15), but there are important differences because there was no previous PTHrP or prolactin. Although recovery is substantial, the bone does not always return to levels that would be expected if DMPA had not been used. Even if the BMD returns to previous values, the skeletal microarchitecture may be damaged. In women undergoing natural menopause, sequential micro-magnetic resonance imaging studies have documented trabecular plate perforations and decreased structural integrity (16). It is not known whether this happens in younger women with either DMPA use or lactation. Potentially, the bone could be lost from surfaces, resulting in trabecular thinning without perforations or structural loss. In this case, the bone strength would be better than if the architecture were destroyed, and recovery would not require reconnection of trabecular elements.

Fracture Risks

Data relating fracture risks to BMD is most abundant in postmenopausal women, when 1 sd decrease in BMD increases the fracture risk about 2-fold. In 10-yr-old children, each sd decrease in bone mineral content, adjusted for size parameters, was associated with an 89% increased risk of fractures in the next 2 yr (17). It is more difficult to determine the relationship between fracture and BMD in healthy young women because the overall fracture rate is so low. One exception is stress fractures in military recruits where the rate was 8.5% in 8 wk. The relative risk for a fracture was 2.26 for each sd decrease in calcaneal ultrasound. In Caucasian women, DMPA use increased the fracture risk by 71%; this became nonsignificant after adjustment for the bone ultrasound (18). The only other study that relates fractures to DMPA was a recent survey of administrative records of medications and fractures in 6773 women with developmental disabilities who were receiving state aid. The DMPA users had an odds ratio for fracture of 2.4 (19). There have not been any large studies of risk of fracture in postmenopausal women who had past use of DMPA.

Other Contraception

In mature women, there is no difference in BMD in those using oral contraceptives. In teenagers, however, this may not be true (20). The Walsh et al. study (5) found a decreased BMD that was not quite significant, and other recent reports suggest that the BMD does not increase to peak values. This may be particularly true with newer doses that contain only 20 μg estrogen. These medicines suppress the midcycle peak of estrogen as well as ovarian androgen secretion, which might result in less anabolic effect on the adolescent skeleton.

The degree of bone loss with DMPA is not severe enough to recommend against it for most women, especially because there is substantial recovery, but physicians must be aware that DMPA could compromise bone integrity of those women at higher risk. Pregnancy itself has many dangers to social and physical wellbeing. These studies should raise a call for development of newer forms of contraception, perhaps a combination of injectable hormones, that will promote skeletal health. Young women need contraception that is convenient, reliable, and safe.