A 55-Year-Old Woman With Osteopenia

DR REYNOLDS: Ms Q was well in 2000 when, at the age of 50 years, she presented for primary care and to inquire whether she should continue taking postmenopausal hormone therapy given her concerns about breast cancer risk. She had begun hormone therapy about a year earlier for health promotion and disease prevention. Her new physician counseled her about her breast cancer risk and her cardiac risk factors and ordered a bone mineral density (BMD) study. Her T score at the spine was –1.4 and –0.6 at the total hip. She continued taking daily conjugated estrogen and medroxyprogesterone acetate until 2001, when she presented with subtle disequilibrium when standing on 1 foot during yoga. Her examination revealed a left-sided subtle ataxia; magnetic resonance imaging of the brain showed multiple tiny infarcts, and further evaluation revealed a small atrial septal defect and a markedly elevated anticardiolipin antibody level. She discontinued her hormone therapy and began taking adjusted-dose warfarin; after physical therapy, her neurological examination returned to normal, although the patient has continued to sense difficulty with balance.

Follow-up BMD testing was performed in 2003 and again in 2005. Her total hip T score had declined from −0.7 in 2003 to −1.0 and her anterior-posterior spine (L1-L4) from −1.9 to −2.3. Her femoral neck T score was −1.9.

Ms Q broke a finger while playing basketball at the age of 40 years. She has an active lifestyle and plays golf regularly but does not always use sunblock. In the winter, she takes her dog for walks, but exercises less than during the golf season. She takes 1500 mg of calcium and a multivitamin daily. She drinks 1 glass of beer or wine nightly; she does not smoke or use recreational drugs. Ms Q works in business and has commercial medical insurance. Her mother had a wrist fracture around the age of 60 years and began taking alendronate after a bone density test when she was in her mid 80s.

Ms Q's past medical history is remarkable for 1 distant episode of nephrolithiasis, episodic benign positional vertigo, and mild recurrent sciatica secondary to a ruptured L3-L4 disk. Her only medications are warfarin, calcium, and vitamin D.

On physical examination, Ms Q is petite. She weighs 106 lb; her height is 61 inches; and her body mass index (BMI) is 20.1. She has normal posture and gait and has normal findings on detailed neurological examination. Her cerebellar testing and tandem walk are normal.

A couple years ago, I noticed that something was just slightly off with my balance. That led us to discover that I’d actually had a couple of mini strokes.

My bone density in the last test had moved to “osteopenia” in a couple of categories. And I actually had had a bone density test a couple of years ago. So I knew that, even from that test, it was likely that the progression was not going to go in a direction as some other people I know who have been told that they could live to be 102 and never have to even have a bone density test. I was disappointed, though, I must tell you, because I do these weight-bearing exercises, I’ve done these stretching and strengthening exercises. I’d done a number of things like taking the calcium where I thought that it would have stayed the same. So when I saw the results, it really made me realize that perhaps this isn't as much in my control as I would like to believe.

It's very unsettling to know that I have a higher potential for getting hurt if I fall. My hopes for the future are that I will remain healthy enough to play golf until I’m one step into the grave, and that my body will stay well enough and flexible enough to allow me to do that because golf to me is the measure of being able to do other things in my life.

I think that my concern about the diagnosis of osteopenia is in the broad sense, that as I get older, this among other things potentially can crop up. And my goal is always going to be how to manage it as best I can. I do not like taking medication at all. It bothers me to have to be held hostage to a Coumadin pill or two every day. Vitamins are easier for me because I view them as being “optional,” sort of, and health-enhancing.

“Osteopenia” is a really neat word, and I know it sits somewhere between “good” and “you don't want to be there.” What I don't know, though, is what it really means to a patient in terms of how they think about mitigating the progression to osteoporosis.

I would like to know if there's anything on the horizon in the research arena for osteoporosis that would allow patients to solve their problem through some wonderful food regimen that would not only satisfy their desire and love of food, but at the same time deal with their bone density issue.

How is osteopenia defined? Is it a disease? What are the risk factors for fractures? What is Ms Q's future risk of fracture? What treatments are indicated for osteopenia? How would they affect her risk for fracture? What are the potential psychological effects of a diagnosis of osteopenia, and what do you recommend for Ms Q?

DR CUMMINGS: Bone mineral density is highly correlated with the strength of bone, and prospective studies in women and men have consistently demonstrated a strong association between BMD and risk of fracture (Figure 1). Bone mineral density of the hip is particularly strongly associated with risk of hip fracture.² The association between decreasing BMD and increasing risk of fracture has no natural biological threshold (Figure 1). Therefore, defining “normal” BMD is necessarily arbitrary.

“Osteopenia” is not a disease. Before the advent of bone densitometry, “osteopenia” was used by radiologists to describe bones that seemed more translucent than normal, and “osteoporosis” was defined by the occurrence of vertebral fracture. The widespread use of densitometry changed these terms. In 1993, a working group sponsored by the World Health Organization met to create consistent definitions of “normal” and “osteoporosis” for BMD testing.³ The group considered using the customary 2 SDs below the mean for young adult women at the femoral neck as the limit of normal. However, this criterion would have led to an unacceptably high 45% prevalence of osteoporosis in women older than 50 years in the United States. A cutoff at −3 or less produced a very low prevalence, out of line with the risk of fractures in postmenopausal women. The conferees settled on 2.5 SDs below young mean BMD in the femoral neck, which produced a 17% prevalence of osteoporosis in the femoral neck among women aged 50 years or older that was similar to the estimated 15% lifetime risk of hip fracture for 50-year-old white women in the United States (written communication, L. J. Melton III, MD, MPH, September 2006). The group added the term “osteopenia” to point out that many women above the threshold for osteoporosis have low BMD and an increased risk of fracture. The upper limit of osteopenia was set at 1 SD below the young adult mean (Table 1). With this upper limit, about half of all postmenopausal white women in the United States would be defined as having osteopenia (Table 1).⁵ There was no clear biological or epidemiological rationale for setting the upper limit at −1.0 (written communication, L. J. Melton III, MD, MPH, September 2006). These terms were intended to allow consistent communication and comparison of prevalence of these conditions across populations and were not intended for use in making treatment decisions. However, in the absence of alternatives, osteopenia and osteoporosis were widely adopted and incorporated into clinical guidelines and reports from densitometry devices.^{6 - 9}

Ms Q's 2005 BMD results put her in the osteopenic range on all measurements; her 2003 results showed “osteopenia” at the spine only. Measurements at 2 sites are only modestly correlated, so their T scores often differ, leading to a diagnosis of osteoporosis at 1 site but not another.¹⁰ A common practice is to use lowest value of femoral neck, total hip, or lumbar spine in assigning a diagnosis.¹¹ This practice will increase the number of women diagnosed as “osteoporotic” or described as having “osteopenia,” but this approach does not improve the predictive value of using only femoral neck BMD to predict nonvertebral or hip fracture.¹²

Osteoarthritis and calcific degeneration of facet joints commonly and artifactually increase spine BMD but not femoral neck BMD, especially after 65 years of age.¹³ A lower spine than hip BMD T score might indicate a relatively increased risk of vertebral fractures.¹⁴ However, the absolute risk of vertebral fractures, based on spine BMD, age, and risk factors, has not been quantified. When there are discrepancies, it is reasonable to focus on BMD at the femoral neck. In Ms Q's case, I would emphasize her femoral neck BMD T score of −1.9 to estimate her overall risk of fractures and make clinical recommendations.

Hearing that she has “osteopenia” is “unsettling,” in the words of Ms Q. Osteopenia is often reported to patients as if it was a diagnosis and patients may believe that they have a disease. Learning that she has bone density that is “below normal” may lead to persistent anxiety.^{15 - 17} Patients should be reassured that they do not have a disease, but this may also represent an opportunity to counsel patients about changing risk factors and lifestyle to reduce their risk of fractures.

“Osteopenia” is defined too broadly to be a useful index of fracture risk. Bone mineral density is only one of many factors that independently influence fracture risk. Age, in particular, is a strong risk factor for most types of fractures and increases fracture risk regardless of BMD. The simple combination of age and BMD can be used to assess a woman's risk of several types of fractures (Figure 2). For example, a 55-year-old white woman with a femoral neck T score of −2 has about an 11% risk of a nonvertebral fracture including less than 1% risk of hip fracture during the next 5 years. Nearly half of all men older than 50 years of age could also be classified as having “osteopenia,”⁵ but their risk of hip and nonvertebral fractures is only about one-third that of women with “osteopenia” of the same age.¹⁸ Similarly, black and Hispanic women called “osteopenic” will have lower risks of fracture than will white women.¹⁹

Estimates of risk of fractures from age and BMD may be improved by considering other clinical risk factors.^{20 - 24} Table 2 lists risk factors that have been associated with an increased risk of hip and vertebral fracture. Besides age and BMD, sex, race, and previous fractures are strong risk factors for most types of fracture. Importantly, a woman who has a vertebral fracture has at least a 4-fold increased risk of having another vertebral fracture, regardless of her BMD.^{22 ,39} Thus, a patient with a vertebral fracture is regarded as having osteoporosis and warrants pharmacological treatment.^{6 - 9} The prevalence of clinically silent vertebral fractures increases with age and lowers BMD. Many dual-energy x-ray absorptiometry densitometers can also be used to detect vertebral fractures. This assessment and bisphosphonate treatment for women with radiographically confirmed fractures may be cost-effective in women older than 60 years with femoral neck T scores of −2.0 to −2.4.⁴⁰ Simple systems for estimating risk of fracture from BMD and risk factors have been published,^{21 - 22} and a system using meta-analyses from many large cohorts is in development.³³

Body weight is a very strong determinant of both BMD and risk of hip fracture.⁴¹ However, weight is not an independent risk factor: once a woman's hip BMD is known, weight (or BMI) adds no information about her risk of hip fracture.^{20 ,41} Therefore, Ms Q's very low weight indicates an increased risk of fracture largely due to her low BMD. Although a history of fracture during adulthood doubles the risk of fractures,³⁴ it seems unlikely that Ms Q's traumatic finger fracture has similar meaning for her risk of fracture. Patients who have had strokes have an increased risk of hip fracture^{30 - 32} due to rapid bone loss and impaired neuromuscular defenses against injurious falls.⁴² Ms Q's tiny infarcts might increase her risk of falling, but her recovery and normal neurological examination suggest that if she has an increased risk, it is substantially less than reported for individuals with hospitalized strokes. In a neurological examination, a patient unable to rise from sitting without using her arms has about twice the risk of hip fracture as a patient who can rise without using her arms.^{20 ,22} Consuming more than 2 alcoholic drinks per day nearly doubles a woman's risk of hip fracture.⁴³ In contrast, drinking 1 or 2 drinks a day, as Ms Q does, is associated with increased bone density^{44 - 46} and does not increase, and might slightly decrease, the risk of fracture.^{43 ,47 - 48} Warfarin blocks the carboxylation of osteocalcin, an essential component of bone matrix, and this has raised concern that prolonged treatment with warfarin would increase the risk of fracture. However, 3 large prospective studies^{49 - 51} have found no increased risk of fracture among women taking warfarin.

Besides estimating risk, it is important to identify risk factors that can be modified or eliminated. Quitting smoking could reduce a patient's risk of both types of fractures by half (Table 2). Less commonly recognized, women with impaired visual function have about a 50% increased risk of hip fracture, and impaired vision can often be improved with appropriate prescription glasses or treatment of cataracts.^{21 ,52 - 55} Thus, testing visual acuity or contrast sensitivity should be a routine part of the assessment of risk of fracture in older patients.

Weight-bearing exercise such as walking (or golfing) may improve bone density of the femoral neck or spine by 1% or 2%^{56 - 57} and reduce the risk of falling and hip fracture.^{20 ,58} Walking and other forms of exercise have been associated with lower risk of heart disease and overall mortality.^{59 - 60} The best type, frequency, and amount of exercise to improve bone strength and risk of fracture is uncertain.

Among elderly women with calcium and vitamin D deficiency, supplementation with calcium reduced risk of hip and other types of fractures.^{61 - 62} However, the benefits of supplementation for healthy ambulatory women, such as Ms Q, are not certain.³⁹ The Women's Health Initiative reported no reduction in fracture risk with 1000 mg of calcium and 400 IU of vitamin D supplementation daily, but 20% to 30% reductions in those who adhered to treatment and among women older than 60 years.⁶³ The supplementation caused kidney stones but only in 4 per 1000 women treated for 7 years. Participants in the trial already had an average baseline calcium intake of more than 1 g daily and the dose of vitamin D may have been too low. A meta-analysis of randomized trials has shown that reduction in fracture risk appears to require supplementation with at least 700 IU of vitamin D daily.⁶⁴

Table 3 lists commonly used prescription drugs that have been approved by the Food and Drug Administration (FDA) for treatment or prevention of osteoporosis. Most practice guidelines agree that pharmacological treatment should be recommended to postmenopausal women who have a vertebral fracture or a T score −2.5 or less at the hip or spine.^{6 - 9} Pharmacological treatment for women with higher levels of BMD is controversial. All approved treatments improve BMD and reduce the risk of vertebral fractures in postmenopausal women with osteoporosis or low BMD. Vertebral fractures are an important cause of disability, particularly among women who already have a vertebral fracture who have a high risk of recurrence.¹⁹ However, nonvertebral fractures account for a substantial share of the disability due to fractures, especially among women who only have low BMD (Figure 3), and only some of the treatments for osteoporosis have been shown to reduce the risk of nonvertebral fractures and hip fracture among women who have a hip or spine T score of −2.5 or less or a vertebral fracture.^{78 - 79} None of the FDA-approved treatments except estrogen have been shown to reduce the risk of nonvertebral fractures in women, such as Ms Q, who have osteopenia. Interestingly, alendronate significantly reduced the risk of nonvertebral fractures, including hip fractures, among women who had femoral neck T scores of −2.5 or less without vertebral fracture but not among women with higher BMD (Figure 4).⁶⁵ A similar pattern was seen for the daily 2.5-mg dose of ibandronate: a significant reduction in risk of nonvertebral fractures only in women with a femoral neck T score of 3.0 or lower, but there was no reduction in risk in women with higher bone density.⁶⁶

Some have developed systems using risk factors to identify women with osteopenia who have a high risk of fracture and who may benefit from pharmacological treatment.^{21 ,82 - 83} Guidelines have not yet been developed based on risk of fracture, and it also remains to be shown that treatment with bisphosphonates significantly reduces the risk of nonvertebral fracture among women with osteopenia plus other risk factors.

Schousboe and colleagues⁸⁴ estimated that, from a societal perspective, 5 years of treatment of postmenopausal women with femoral neck T scores above −2.5 without a vertebral fracture would cost $70 000 to $332 000 for each quality-adjusted life year (QALY) saved by prevention of fracture. At the commonly used benchmark of $50 000 per QALY, treatment would not be cost-effective. However, they estimated that if a woman had other risk factors, such as sustained use of corticosteroids, that at least doubled her risk of vertebral fracture, treatment would be cost-effective.

Ms Q's risk of fractures can be estimated from her age and femoral neck BMD (Figure 2). During the next 5 years, she has about a 3% risk of developing vertebral fracture detectable by spine films, of which one third would come to clinical attention. Her 11% risk of a nonvertebral fracture over the next 5 years includes a very low (<1%) risk of hip fracture. Her risk factor profile does not substantially change these estimates. Her history of small strokes and tripping falls may increase the risk of nonvertebral fracture a little, while her regular recreational exercise and regular habit of drinking wine slightly decrease her fracture risk. The risk due to her low BMI is reflected in her hip BMD and, as noted earlier, neither her history of a traumatic finger fracture nor use of warfarin should influence her risk.

Exercise is perhaps the most important recommendation and has multiple benefits. Ms Q plays golf, an ideal exercise if she walks rather than riding a cart. She takes 1500 mg of calcium and a multivitamin (which usually contain 400 IU of vitamin D or less). The benefit of calcium and vitamin D supplements for a healthy active middle-aged woman is not clear. As the average diet of postmenopausal women in the United States is about 700 mg of calcium per day, 500 mg is generally all that is needed to achieve an intake of 1 g daily. Reducing her supplemental calcium dose might also reduce her small risk of kidney stones. A meta-analysis found that supplementation with vitamin D reduced risk of falls by about 20% in older women,⁸⁵ but it is not clear that this is applicable to healthy active women such as Ms Q. Nevertheless, 800 IU of vitamin D daily carries no known risk and is a reasonable recommendation for Ms Q.

Between 2000 when she was taking estrogen and 2005 after she had discontinued estrogen therapy, her BMD decreased from a T score of −1.4 to −2.3 in the spine and −0.6 to −1.0 in the total hip site. The change, most notably in spine BMD, may be due to discontinuation of estrogen therapy. Discontinuation of estrogen therapy may be followed by several years of rapid bone loss, particularly in the spine.^{79 ,86 - 87} If this is the reason for the change in BMD, then the rate of loss should slow. On the other hand, low weight is correlated with low levels of estradiol, which leads to somewhat faster and persistent bone loss.⁸⁸ Seemingly large changes between 2 measurements can occasionally be due to variability in the measurement, and BMD often increases on repeat measurement.⁸⁹ I would test again in 1 or 2 years; persistent rapid loss would tip the scales toward pharmacological treatment.

Experts would differ about whether Ms Q's risk of fractures warrants taking a prescription drug to decrease her risks. Nonvertebral fractures constitute her greatest risk. Estrogen is the only drug with evidence for reducing the risk of those fractures in women with osteopenia, but is contraindicated in Ms Q because it increases the risk of stroke. Bisphosphonates and raloxifene could be considered, but there is no evidence that they would reduce that risk (Table 3). She does not like taking medication and I would not recommend it. I would reconsider this recommendation if a strong risk factor, such as weight loss or recurrent falling, developed. I would reassess the rate of change in BMD in 1 or 2 years.

Pharmaceutical companies are testing many new drugs for osteoporosis. Fewer nutritional approaches are under study. High levels of homocysteine indicate a substantially increased risk of hip fracture.⁹⁰ Folate and vitamin B₁₂ supplementation reduces homocysteine level and may reduce the risk of hip fractures.⁹¹ Diets high in animal protein and relatively low in fruits and vegetables lead to chronic mild metabolic acidosis that may promote bone loss.⁹² Studies are under way to determine whether restoration of normal acid-base balance slows bone loss.

The term “osteopenia” is losing favor and there is growing dissatisfaction with the use of T scores.^{93 - 94} In the near future, these terms will probably be replaced by estimates of a patient's risk of fractures as a basis for making recommendations about treatment.³³

QUESTION: Are risk assessments available on the Internet for people to assess their risk of fracture?

DR CUMMINGS: In the next year or two, a system of assessing risk, developed under the auspices of the World Health Organization,³³ may be widely available, perhaps on the Web, for estimating risk as a part of bone densitometry reports. Black and colleagues published a validated system for estimating fracture risk in elderly women.²² The simple charts in Figure 2 can also be used to approximate risk based on hip BMD and age, but these are not yet available on a Web site.

QUESTION: Whom should we choose for BMD measurement among men and women? Is screening bone densitometry a good way to spend the nation's resources?

DR CUMMINGS: There is a consensus in the United States that it is worthwhile to recommend BMD testing to women who are age 65 or older or younger postmenopausal women who have a strong or multiple risk factors for fracture.^{6 - 8 ,95} Men have a substantially lower risk of fracture than women and, therefore, warrant a more conservative approach. However, evidence and guidelines for men are very limited. I think it is reasonable to recommend densitometry to men age 65 or older who also have least 1 strong risk factor, such as a previous fracture.⁶

If a younger or lower-risk postmenopausal woman requests BMD testing, I would order the test. The problem is that if hip and spine BMD are measured, she has at least a 50% chance that she will be told that she has osteopenia, which may be taken to mean that she has a disease that needs treatment. With an understanding that osteopenia is not a disease that needs a drug, counseling about her absolute risk may allay inappropriate anxiety and perhaps motivate beneficial changes in lifestyle. Ms Q is in this group; younger than age 65, her physician ordered the BMD test at a time when it might have helped to inform a decision about taking hormone therapy.

QUESTION: You talk about the importance of looking at risk. How do you weigh mild osteopenia by bone density in a patient who has multiple risk factors? We see that frequently in our internal medicine practice: patients with falls, low body weight, stroke, and just mild osteopenia.

DR CUMMINGS: If “mild osteopenia” means a hip BMD T score above −1.5, then I would not recommend treatment unless the patient had other very strong risk factors, such as a recent stroke. If you know the patient's BMD, then her weight is not an additional risk factor. It is reasonable to consider pharmacological treatment in a patient who has a BMD T score between −1.5 and −2.5, if her age and other strong or multiple risk factors indicate a high risk of fracture. However, recommendations should also take into account the efficacy for reducing the risk of nonvertebral fracture for the treatment you are considering and the patient's preferences about taking medications to reduce her risk.

QUESTION: You said everybody at 65 should get a bone density scan, and I wonder if you have a cutoff at the other end? When is it no longer cost-effective to get a bone density scan? We see a lot of healthy 90-year-olds in our practice.

DR CUMMINGS: The risk of hip fracture is very high after age 80 and, in women with osteoporosis, treatment can significantly reduce the risk of hip fracture within 2 or 3 years.^{67 - 68} Screening for osteoporosis and treating those with a hip BMD of −2.5 or less appears to be cost-effective for healthy elders even after 85 years of age.⁹⁶ However, screening should not be done if the patient's life expectancy is so short that you would not consider treating to reduce her risk of fracture.

QUESTION: One issue is the potential long-term dangers of bisphosphonates. Are you recommending a specific time limit for treating with bisphosphonates?

DR CUMMINGS: There have been concerns that very long-term suppression of bone remodeling by bisphosphonates might weaken bone.⁹⁷ However, continuing alendronate for up to 10 years is probably safe.^{98 - 99} The FIT Long-term EXtension [FLEX] Trial randomly assigned women who had taken alendronate for 5 years to take placebo (stop alendronate) or continue treatment for 5 more years.¹⁰⁰ During the next 5 years, those who stopped had the same rate of nonvertebral and radiographic vertebral fractures as those who continued. Those who continued had a 50% decrease in the relative risk and about a 3% decrease in the absolute risk of clinically recognized vertebral fractures over 5 years. It is reasonable to consider at least a 5-year holiday for patients who have been taking alendronate for at least 5 years but continue treatment in those with a high risk of vertebral fractures, such as those who have previously had a vertebral fracture.

QUESTION: And many of the patients that we take care of in this situation will call within a few months and say, “May I have another bone density, because I want to know whether I’m still losing.” What would be the appropriate timing for the follow-up bone density?

DR CUMMINGS: Measurements of bone density are quite precise but still have noise. In general, changes in BMD between 2 tests must exceed at least 4% to 5% to provide confidence that the change is real. With longer intervals and more than 2 measurements, you get a more reliable picture of the real change in bone density.

Doctors often recommend follow-up measurements of BMD every 2 years, a frequency that is often covered by insurance plans. However, the interval should be based on a patient's current BMD, the likely rate of change, and the level at which treatment would change. On average, older postmenopausal women lose less than 0.5% BMD per year at the femoral neck.¹⁰¹ A change of 1 SD (1 T score unit) represents about 10% change in BMD. The rate is more rapid in the first few years of menopause and after age 75.¹⁰¹ If a postmenopausal woman is losing BMD at twice the average rate, it would take about 10 years for her femoral neck T score to decrease 1 unit. Therefore, if treatment would be recommended at a T score of −2.5, it is reasonable and conservative to recommend that a patient with no other risk factors and a T score of −1.5 repeat the test in 5 to 10 years.

In Ms Q's case, I would test again within 2 years because her BMD has been decreasing relatively rapidly since stopping estrogen therapy.