SCREENING AND DIAGNOSIS OF BONE HEALTH IN
POSTMENOPAUSAL WOMEN: PREVENTION
Deborah Manzi-Smith, MD
Ruben Alvero, MD
IDENTIFYING THE PROBLEM
Osteoporosis is a common disorder that places a large medical and economic burden on the United States healthcare system. As a result of the "aging of America", the direct and indirect costs of managing fractures due to osteoporosis are expected to rise dramatically. A large portion of postmenopausal women seen in primary care practices have significantly reduced bone mineral density (BMD) and an increased risk for fractures.1 The clinical importance of this is not the disease itself but the fracture risk associated with the disease. According to conservative estimates, a 50-year-old Caucasian woman has a remaining lifetime risk of 40% for hip, vertebra, or wrist fractures.2 This morbidity has considerable medical, financial, and social ramifications, especially in light of the fact that the mortality rate is 20% higher within one year of a hip fracture.3 It has been estimated that the annual cost to the health care system for osteoporosis related fractures is $17 billion, with the estimated cost of $40,000 per patient for a single hip fracture.4
Although osteoporotic fractures are an important cause of morbidity and mortality, it is recognized as a preventable disease. Addressing risk factors early may help prevent devastating postmenopausal bone loss.5 Despite effective primary and secondary preventions, osteoporosis is often under-diagnosed and therefore under-treated. These clinical practice guidelines are intended to summarize the best evidence to help health care providers make clinical decisions about osteoporosis diagnosis and prevention.
DEFINING THE PROBLEM
Osteoporosis was defined in a 1993 consensus conference as "a systemic skeletal disease characterized by low bone mass and microarchitectural deterioration of bone tissue with an increase in fragility and risk of fracture".6 Many terms are used to describe the density of bone. Bone mass refers to the amount of bone in the skeleton; however currently, no technology produces a measurement called bone mass. Another frequently used term is Bone Mineral Density (BMD). BMD is defined as the average concentration of minerals per unit area of bone with results reported as T-scores and Z-scores. A T-score represents the difference in standard deviations of the patient's BMD compared to a young, healthy adult of the same sex and race. WHO classifies a T-score lower than -2.5 as osteoporosis and a T-score between -1.0 and -2.5 as osteopenia.6,7 Although less frequently used, a Z-score may also be utilized. The Z-score represents the difference in standard deviations of a person's BMD compared to persons of the same sex and age. One should suspect secondary causes of osteoporosis when the Z-score is lower than -1.5 SD.8
IDENTIFYING THOSE AT RISK
In women, peak bone mass is achieved by the second decade and in normal women this peak bone mass is maintained until the time of menopause. The determinants of the severity of the osteoporosis at the time of menopause are influenced by poor bone mass acquisition during adolescence and accelerated bone loss during perimenopause/menopause. It only makes sense that women who never reach their peak bone mass will have lower than normal bone mass in their early thirties. These women will have less bone to lose, and therefore are at high risk for the osteoporosis when menopause begins (Table 1).
Many factors are known to affect bone mass. These include the "non-modifiable" risk factors such as family history, female gender, age, ethnicity, and history of fractures and the "modifiable" risk factors, which include cigarette and alcohol use, nutritional status, estrogen status, and physical fitness status.
The Non-modifiable Risk Factors
History of Fractures
The Modifiable Risk Factors
Smoking and Alcohol Use
Risk for Falls
SCREENING FOR BONE LOSS
Who Should Be Selected for Screening?
Guidelines for Screening
History and Physical Exam
Also useful are the biochemical markers of bone turnover. Bone turnover is a normal process that maintains bone strength. During bone remodeling the osteoblasts release factors into the blood and measurement of these factors reflects the rate at which the osteoblasts are working (rate of bone formation). Osteoclasts break down bone and release degradation products into the blood stream that are eventually cleared by the kidneys. Measurement of these breakdown products is a reflection of osteoclast activity. These biochemical markers of bone turnover include bone specific alkaline phosphatase, procollagen I carboxyterminal propeptide (PICP) and osteocalcin which are indices of bone formation. The pyridinolines, deoxypyridinolines and the type I collagen telopeptides (Ntx or Ctx) are indices of bone resorption. These markers are of value in estimating bone turnover rates and may help identify "fast bone losers".33 the combination of a BMD and biochemical markers may help us better identify which patients with osteoporosis are at greatest fracture risk. It is known that a person with a low BMD and high resorptive markers has double the risk for fracture compared to persons that have low BMD and normal resorptive markers.34 Biomarkers may be useful in monitoring response to osteoporosis therapy. In a severely osteoporotic patient, serial biochemical markers will provide information about a patient's response sooner than serial DEXA scans.
BMD measurement of the peripheral skeleton is usually accomplished by quantitative ultrasound (QUS). Testing by QUS of the heel is less expensive than DEXA and has become popular. QUS of the heel has been shown to predict hip fractures and non-vertebrae fractures as well as DEXA; however, if QUS is used for screening, confirmation of the diagnosis of osteoporosis is recommended with DEXA.35 Before QUS can be considered as a replacement for central DEXA, large prospective trials must be undertaken to demonstrate that it is at least as good as the DEXA technology for fracture prediction and treatment follow up.
PRECISION IN BMD TESTING
Evaluating changes in BMD over time can determine the rate of bone loss. However, it is important to note that the average rate of bone loss in postmenopausal women is 0.5%-2% per year and most therapies increase BMD by 1-2% per year. Given these relatively small percentages, only a precise instrument will detect short-term changes. It is important to keep in mind that the short and long term variability of an instrument should be taken into account when assessing BMD changes. The DEXA scan has about a 2% short-term variability in assessing BMD.
DETERMINING FACTORS FOR INTERVENTION
As is generally true in other areas of medicine, prevention is the most effective intervention for this disorder. Decreasing fracture risk should start in childhood and adolescence with appropriate dietary and exercise habits.36 Smoking and caffeinated beverages should be avoided and nutritional sources of calcium and Vitamin D may need to be enhanced with supplements. By establishing adequate bone mass in the years prior to peak levels, it is hoped that the patient will have an adequate cushion as she enters the hypoestrogenic postmenopausal period.
For those women already in their menopausal years, the American Association of Clinical Endocrinologists (AACE) suggests that treatment should be initiated in patients with low-trauma fractures, T scores of -2.5 and below, T scores of -1.5 and below with risk factors present and where preventive measures are ineffective.37 Other risk factors the AACE recommends be evaluated include history of hip fracture in a first-degree relative, weight loss and low body weight, cigarette smoking, increased likelihood of falling, tall stature, high bone turnover and advancing age. The North American Menopause Society (NAMS) further specifies that women over the age of 65 should be evaluated with BMD and possibly treated regardless of any additional risk factors.38
As has previously been noted, however, the BMD is often unknown since universal screening for women is not recommended because of the poor predictive power of the study in predicting an individual's absolute risk of fracture. Correct identification of a population to evaluate is especially important in light of the fact that it is estimated that only a third of osteoporosis cases in the US have been identified and one seventh of such cases are being treated. A meta-analysis reviewing over 90,000 person-years of observation and over 2000 fractures has confirmed this observation and noted significant overlap in BMD between patients who had a fracture and those who didn't. 39 While the use of clinical risk factors would appear to be a reasonable approach, the attributable risk based on these alone appears to be quite low.40
Given the poor sensitivity for any given specificity in BMD and clinical risk factors in targeting risk adequately some have suggested that diagnostic thresholds be used to evaluate patients. This is a type of risk scale using BMD and clinical factors, with the resulting nomograms used to individualize patient care. The use of these therapeutic thresholds lend themselves readily to cost-effectiveness analysis and a ten year cumulative fracture risk has emerged as a touchstone for intervention in several studies. Not surprisingly, age emerges as a critical factor in such longitudinal studies. Sweden has among the most thoroughly evaluated post-menopausal populations in the world. Studies from Sweden suggest that a 10% 10 year risk of fracture is an acceptable threshold for treatment.41 Alternatively, achieving age 70 is another cost-effective milestone that appears to justify treatment in this Nordic population.13 An opportunity for error exists in that the rate of bone loss as patients age is not likely to be linear and therefore risk stratification may underestimate fracture thresholds in older patients.42 Of course intervention thresholds are additionally affected by other assumptions that are made in the model created. Variables such as cost of the intervention, effectiveness of treatment, inclusion of all osteoporotic fractures and certainly age of the patient all had a significant impact on justification for treatment in yet another study from Sweden. 43 The authors of this study additionally cautioned about the external validity of their results on a population outside of Sweden. It should also be remembered that certain fractures, such as those of the foot and ankle, are very weakly associated with low bone mass and therefore should not be included in a threshold model.44
The concept of a threshold effect is not universally accepted and many have argued that, given the magnitude of osteoporosis and the cost involved in failing to treat someone who goes on to have a fracture, watchful waiting is not appropriate, even in relatively young postmenopausal women with osteopenia.45 Because clinical trials target high risk patients, little data exists on the benefits of intervention in patients with osteopenia with regard to fracture risk. Citing the beneficial impact of tightly correcting blood pressures, Law and Wald argue that reducing all risk variables as much as possible cannot help but reduce the incidence of fracture in postmenopausal women.46 Arguing for treating the large mass of patients at any risk, they note that there is a continuous relationship between risk factors and the disease they cause at all levels and not just those at a critical value. This is especially true in light of the fact that some researchers have noted that even taking into account all of the conventional risk factors for hip factor, the attributable risk of these factors was only 70% in the populations studied.47 While they correctly argue that there is no true threshold and a significant overlap exists in bone mass between those who go on to fracture and those who don't, the argument does not take into account cost and side-effect profile of the medications given.
In addition to identifying those whose risk threshold should trigger treatment, the type of therapy should also be tailored to the cause of the osteoporosis or increased fracture risk. Human bone is approximately 80% cortical and 20% trabecular but the latter is much more sensitive to metabolic influences that will lead to rapid bone loss and significant bone loss at the predominantly trabecular spine may be more amenable to early intervention.48 Additionally, identification of fractures from external sources such as falls rather than low trauma injuries would further target the intervention.41 Since secondary causes of osteoporosis may be associated with large losses in a short period of time, the history should at the very least evaluate if any of these secondary causes exist (Table 4). This is especially true in light of the fact that it is estimated that almost a third of osteoporosis cases have a secondary cause as at least part of the source of the disorder. 49
Certain subpopulations of postmenopausal women deserve special attention. The frail elderly are at especially high risk from several perspectives. Since these women are by definition older, the overall risk of low bone mass is exceptionally high. In addition, the diminished capacity to perform activities of daily living make them more vulnerable to malnutrition and weight loss with consequent further deterioration due to inactivity and decrease in substrate necessary for bone strength.50 Attempts at modifying the patients' interaction with her environment and introduction of protective devices such as hip protectors may in part diminish the risk of low trauma fracture. Removal of hazards, improvements in lighting, improved activity levels and enhanced exercise can benefit these patients. In addition, maintenance of body weight, walking for exercise, treating visual impairments and avoidance of long-acting benzodiazepines can further ameliorate risk.51 Paradoxically, a Canadian study examining community dwelling patients older than 65 years revealed that an increase in weekly activities actually increased the risk of injurious falls.52 Although this study did not specifically address osteoporotic fracture risk, this clearly suggests that guidance in the nature of activity is important to enhance the benefit without incurring excess risk. Unfortunately, the data supporting the efficacy of such interventions is limited and compliance in an elderly population is questionable at best.
An opportunity for intervention that is frequently missed is in those patients who sustain their initial osteoporotic fracture but go untreated even when seen by their provider.53 Since arguably this is the highest risk population and there is a significant incidence of repeat fracture in the first year after an initial vertebral fracture54, a program of education and heightened awareness, particularly among orthopedic surgeons and emergency physicians would presumably go a long way to achieve cost-effective treatment. This fact is underscored in recent studies which have shown that only 20% of women treated with minimal trauma fractures at the hip, spine and wrist were subsequently treated for osteoporosis and there was actually an inverse correlation with age and likelihood of therapy.55
Specific attention to neuromuscular function can also be expected to reduce the risk of fracture. In a well-done study in a large cohort of women greater than 75 years of age, poor vision again was cited as a significant risk-factor for hip fracture. Neuromuscular function as assessed by tandem heel walk and gait speed were also significant in predicting an injurious fall.29 When the combination of poor vision or neuromuscular compromise and low BMD was assessed, the risk of hip fracture was 29 per 1000 woman-years. Presence of just one or the other factor predicted 11 fractures per 1000 woman-years. Absence of fall risk factors or low BMD set the incidence at 5 per 1000 woman-years. Clearly the potential exists for intervention in the patient environment that can clearly decrease the morbidity and mortality of hip fracture.
Another potential intervention is the reduction of cigarette smoking. Duration of smoking appears to be more important than the dose and interval of smoking cessation is also inversely proportional to the risk of hip fracture and current, postmenopausal smokers had the highest incidence of such events.56 The same study found that concurrent alcohol use had a weak ameliorative effect on the population studied. While alcohol should not be encouraged as a specific intervention to reduce osteoporotic fracture, women over 65 appeared to have a dose-response improvement in BMD in one population studied.57
As the most common cause of low bone mass due to secondary causes, glucocorticoid use is particularly important. Glucocorticoids have an adverse effect on osteoblasts, thereby reducing bone formation and increases osteoclastogenesis as well as reducing osteoclast apoptosis.58 There is also a decrease in calcium absorption in the GI tract and increase in renal calcium excretion. The net effect is significant BMD compromise if no intervention is made early in the course of treatment. The American College of Rheumatology stresses the importance of prevention and recommends the institution of treatment in women beginning the prednisone equivalent of > 5 mg/day for > 3 months.59 Treatment should be continued for as long as the patient is receiving glucocorticoids. The treatment used in postmenopausal women on glucocorticoids should consist of lifestyle changes designed to diminish osteoporotic risk as well as pharmacologic treatment with bisphosphonates (Table 5).
In summary, consensus exists with regard to the value of intervention in more extreme cases and age in particular is cited as a significant contributor to risk of osteoporotic fracture. Controversy exists where the risk is more modest. While intervention is clearly indicated in patients where osteoporosis has been identified by radiologic means, in the absence of universal screening, providers have to rely on clinical judgment and the presence of other risk factors. As has been demonstrated, however, even the combination of low BMD and other risk factors does not entirely predict fracture and the practitioner is left to ponder the individual risk. At the very least, increased sensitivity to such frequently overlooked factors such as previous fracture, glucocorticoid use and age should be beneficial in preventing many fractures. Recognition of visual and neuromuscular compromise should further allow non-pharmacologic interventions to reduce morbid injuries. In the end, the provider is left to use broad guidelines suggested by national organizations (Table 6) with subsequent individualization on the basis of the costs that patients (and society) are willing to pay to derive a particular benefit.60
NUTRITION, CALCIUM AND VITAMIN SUPPLEMENTATION
In humans, 99% of total body calcium is found in the bones, 0.9% is intracellular and 0.1% is in the extracellular fluid where it is responsible for many physiologic processes. Calcium requirements vary significantly through a woman's life and are affected by enhanced utilization such as occurs during growth and development in childhood as well as pregnancy and by variations in gastrointestinal absorption. While calcium homeostasis is balanced during much of adult life, change in the balance towards absorption begins at the perimenopause under the influence of diminished ovarian steroids.
As osteoporosis can be induced by calcium deficiency experimentally, it is plausible that these changes can modify the complex web of events that impacts bone deposition and absorption. There are many nutritional factors that impact the amount of available substrate for bone remodeling. It has long been felt that protein intake is directly related to the amount of calcium excreted in the urine and therefore has been associated with low BMD. Epidemiologic data such as the high rate of osteoporosis in those populations consuming a Western diet seem to support such a notion although a recent study using the Framingham database disputes this claim and suggests that there is low BMD in those consuming low amounts of protein in their diet.61, 62 Amounts of dietary sodium are directly related to the amount of calcium removed by the kidneys. Vitamin D levels, which are known to decrease in tandem with changes in skin architecture and decreased outdoor activities in older individuals, also strongly impact the amount of available calcium.
Since peak bone mass is the primary determinant of BMD in the early stages of bone loss, it has long been argued that improving preventive measures such as calcium intake early in life places the early menopausal woman in an advantageous position with regard to the initial impact of hypoestrogenism. Although many of the early studies were small and had variable results with regard to the impact of premenopausal calcium intake, more recently researchers have demonstrated reasonable support for this concept. 63 In comparing teenage intake of calcium with supplementation in women in their 30s, one study found the former to be a stronger predictor of BMD in the hip but not the spine or the distal forearm. A dose effect was also found with an increase in calcium from 800 mg to 1200 mg daily intake associated with a 6% increase in hip BMD. Overall, adequate calcium intake in the teenage years was a stronger predictor of BMD than current use and the effect was enhanced in physically active subjects.
Later in life, the impact of calcium supplementation is greatest beyond five years after the menopause. It is difficult for calcium replacement in the absence of antiresorptive therapy (HRT or bisphosphonates) to positively affect bone homeostasis during this time period because of the overwhelming effect of hypoestrogenism. 64 Beyond this early time period, however, calcium supplementation, particularly in conjunction with supplemental vitamin D, there is a modest impact on reducing bone loss in the femoral neck, spine and total body.65 More importantly, there appears to be a reduction in non-vertebral fracture risk in those treated with calcium and vitamin D supplements and a Cochrane Database review supports this conclusion. In a review of fifteen trials with a total of 1806 subjects a reduction of bone loss was apparent but the impact on fracture risk reduction was less evident, possibly affected by small numbers in the available trials.66 Although beyond the limits of this review, antiresorptive therapy such as with estrogen therapy appears to potentiate the effect of calcium on BMD in postmenopausal women. In a review of 20 studies, the effect appeared dramatic at times with enhancement ranging in the 3-fold range. Although the impact of anti-resorptive therapy was greatest in women greater than five years beyond the menopause, an effect was seen in peri- and postmenopausal women as well.67
Calcium is probably best replaced in the diet along with other important nutrients. Nevertheless, many women find it difficult to completely meet their requirements with dietary sources alone. Dairy products, particularly those fortified with calcium and vitamin D, are the most efficient means of obtaining calcium (Table 7). Green leafy vegetables also contain calcium, albeit in lower concentrations and they may also reduce bioavailability since they contain oxalates, which may compromise GI absorption.8, 64
As previously noted, calcium requirements vary throughout a woman's lifetime and also depend on the hormonal milieu (Table 8). Recommendations take into account the efficiency of GI uptake at the given patient age as well as availability of other cofactors such as vitamin D and likelihood of exposure to sunlight which is a key ingredient in vitamin D metabolism. If supplementation is required, the patient should take no more than 500 mg of elemental calcium at each dose since this improves the efficiency of GI uptake. Taking the supplements between meals also enhances efficiency and eliminates the possibility that certain agents, such as oxalates, inhibit uptake. A notable exception may be in older patients who have decreased gastric acid secretion since the most common supplement available (calcium carbonate) is less bioavailable if there is no gastric acid.68 Calcium citrate does not depend on gastric acid for activation but is more costly and has a lower percentage of elemental calcium. Finally, it is important to note that it is the amount of elemental calcium that is important and different formulations contain varying amounts of this form of the supplement (Table 9). While calcium carbonate is the least expensive formulation, it can also inhibit iron absorption and is therefore a problem in women with iron deficiency anemia taking this supplement. Calcium citrate does not have this problem and actually enhances iron uptake.
The metabolites of vitamin D, principally 1, 25-dihydroxyvitamin D, drive the active uptake of calcium in the small intestine and the colon. Sources of vitamin D include sun exposure, dairy products, particularly those fortified, fatty fish, cod liver oil and supplements. Although the current recommendations from the Standing Committee on Dietary Reference Intakes is for 400 IU/day in women aged 51-70 and 600 IU/day in those older than 7064, there is considerable evidence that these recommendations may be too conservative and many researchers feel that 600-800 IU/day is more appropriate for all menopausal women.68, 69 The basis for this recommendation is the recognition that the majority of these women are high risk on the basis of low exposure to sunlight due to inactivity. In any case, the majority of studies looking at both calcium and antiresorptive therapies were conducted with a baseline of vitamin D supplementation in this range. In their zeal to achieve adequate levels of vitamin D, however, women should not overcorrect by doubling their vitamin dosing since excessive amounts of vitamin D are associated with hypercalciuria and the accompanying excess of vitamin A may be associated with increase in fracture risk.70
An intriguing and novel area of research in vitamins and bone health involves the intake of vitamin K. Classically, vitamin K has been critical in the coagulation cascade but it now appears that vitamin K, as a cofactor the carboxylase enzyme that is instrumental in conversion of glutamyl to ã-carboxyglutamyl, is important in osteocalcin carboxylation.71 In addition to this positive influence on osteoblast activity, vitamin K may also be instrumental in calcium retention at the level of the kidneys. Epidemiologic data supporting this beneficial effect comes from the Nurses' Health Study. Women in the lowest quintile of vitamin K intake had higher risk of hip fracture than women in quintiles 2-5 and this risk remained in spite of adjusting for calcium and vitamin D intake.72 Lettuce and other green leafy vegetables are good sources of vitamin K. In light of this and other evidence, the Institute of Medicine has increased the dietary reference values for vitamin K to 90 µg/day for women.
Magnesium is widely used as a dietary supplement and there has been considerable interest in its impact on calcium absorption. Unfortunately, there is currently no evidence that magnesium significantly impacts calcium homeostasis in the human.73
Low bone mass is a significant health problem in postmenopausal women. In addition to the morbidity and mortality, there is significant financial cost to the health care system and to the society. Prevention can reduce many of the fractures resulting from this osteopathy and the encouragement of a healthy lifestyle with exercise, good nutrition and decreased smoking can reap rewards in other areas beside bone health. Intervention remains somewhat controversial and at the discretion of the provider and patient but a heightened awareness of risk factors such as age, prior osteoporotic fractures and secondary causes remain important. Future research should continue to focus on those patients with less severe bone loss to assess the impact of treatment on their risk of future fracture.
Modify lifestyle risk factors for osteoporosis
Advanced Reproductive Medicine
Detailed Directions to Our Offices