Diagnosis

There are several diagnostic approaches that may be used to identify postmenopausal women who may have osteoporosis or are at risk for developing osteoporosis. These evaluations include:

• risk factors and Simple Calculated Osteoporosis Risk Estimation (SCORE)
• clinical features
• laboratory tests (e.g., biochemical markers of bone turnover)
• radiologic techniques to assess bone
• measurement of bone mineral density (BMD) by noninvasive techniques.

Risk Factors

From the medical history, the physician can evaluate the risk factors and any associated conditions; important factors include menstrual history and menopausal status, family history, medication history, and concomitant disease states. Current and previous medications are reviewed with special attention to glucocorticoid and hormone use. Note that nutritional and dietary factors do not simply refer to current dietary intake of calcium; consideration should be given to lifelong patterns of calcium intake, mineral and vitamin supplementation, and clues related to malabsorption of calcium, nutritional deprivation or anorexia, and vitamin D deficiency. Environmental factors, such as sunlight exposure and current and lifelong level of exercise, are also important in the evaluation.

Simple Calculated Osteoporosis Risk Estimation (SCORE) [Lydick et al., 1998] is a simple test, based only on risk factors, that was developed to help identify postmenopausal women age 50 and older who are likely to have low bone mass and who should be evaluated further for osteoporosis. The goal of the questionnaire is to improve patient selection for bone densitometry and thus increase the cost-effectiveness of screening for osteoporosis.

The questionnaire was developed so that patient information could be easily obtained during a study of community-dwelling perimenopausal and postmenopausal women of different races and ethnicity, ages 45 and older [Lydick et al., 1998]. Women answered about 60 questions on factors suspected of being associated with osteoporosis. The 60 questions were chosen based on an extensive review of the osteoporosis literature. An analysis of the study data led to the development of a simple form with six scored questions.

SCORE can be used to assign postmenopausal women to one of two groups:

• women unlikely to have low BMD and, therefore, probably not currently candidates for bone densitometry
• women likely to have low BMD and, therefore, probably appropriate for further evaluation by BMD testing.

According to the scoring system women with a final score of >/=6 should be evaluated further for osteoporosis.

As illustrated in the SCORE sheet below, the factors most strongly associated with low BMD were determined to be race, rheumatoid arthritis history, fracture history, weight, age, and estrogen use. The scoring system was validated in a different group of 207 postmenopausal women and was found to have a sensitivity of 90% and a specificity of 40%. The scoring system correctly classified about 62% of the women in the validation cohort (90% of women with low BMD and 40% of women with non-low BMD) [Lydick et al., 1998].

Sample SCORE Sheet

Click on image for larger version.

Data from this study suggest that, in a population of 1000 women similar to those in this study (assuming 38% prevalence of low bone mass), using SCORE as a pretest set at 90% sensitivity (and a specificity of only 40%), densitometry could have been avoided in about 30% of the women with low BMD and only 10% of women with low BMD would have been missed. Also, 62%, or 200, of the 341 women with osteopenia (BMD between 1 and 2 standard deviations (SDs) below the young adult mean) would have been identified.

Age is recognized as an independent risk factor for osteoporotic fracture and is a key consideration in predicting the patient’s rate of bone loss. Bone mass increases progressively during childhood and adolescence, reaching a peak in early adulthood (ages 25-35). Gradual bone loss, or age-related bone loss, then begins in both men and women (at a rate of about 0.3-0.5% per year). At menopause, women begin to experience accelerated bone loss. The rate of bone loss is most rapid (3-5%) in the first few years (5-7 years) after menopause [Levinson and Altkorn, 1998]. Generally, bone mass declines at a lower rate thereafter, although the rate of bone turnover may remain elevated for the entire duration of postmenopausal life. Over their lifetimes women lose about 30-50% of their total bone mass, whereas men lose about 20-30% of their total bone mass [ Riggs et al., 1981]. Other factors that are associated with an accelerated rate of bone loss include immobilization, corticosteroid therapy, and treatment with excessive doses of thyroid hormone.

Low bone mass is the most important single determinant for assessing the risk of experiencing an osteoporotic fracture. Although it is not the sole determinant of fracture risk, its value in predicting fracture risk has been found to be better than that of cholesterol measurements in predicting the risk of heart disease or blood pressure measurements in predicting the risk of stroke [Kanis, 1994].

Clinical Features

Osteoporosis is defined on the basis of low bone mass; thus, early in the disease process there may be no clinical symptoms. Bone loss associated with osteoporosis is generally an asymptomatic process. Making the diagnosis can be challenging for the health care provider. Even vertebral fractures may be asymptomatic and go undetected. With the development of painful fractures, the disease usually becomes symptomatic and more easily recognized clinically.

Many factors contribute to the development of osteoporosis, including:

• failure to develop sufficient bone mass in young adulthood
• accelerated bone loss due to estrogen deficiency at menopause
• age-related bone loss
• genetic contribution to bone health (family and personal history of fragility fracture)
• long-term inadequate calcium intake
• defective intestinal absorption of calcium.

A complete medical history and thorough physical examination are the foundation for the accurate diagnosis of osteoporosis and the effective management of the disease (see table below). The physician, through a patient interview and a review of previous medical records, learns about the patient’s medical history. Present medical complaints, symptoms, family history, and physical appearance are key pieces of information obtained as part of the medical history.

Components of a Physical Examination in the Evaluation of Osteoporosis

Determination of patient’s height and weight

Examination performed in good light with patient undressed and standing erect

Examination of the back to determine degree of spinal deformity and any areas of vertebral tenderness

Assessment of muscle strength for muscle atrophy

Determination of range of joint motion and flexibility

Assessment of signs and symptoms associated with estrogen deficiency (e.g., thinness of vaginal tissues, hot flashes, history of hysterectomy)

Examination of feet, observation of gait, and assessment of balance

Performance of a neurologic examination

Performance of a general systemic examination

Signs and Symptoms

Although osteoporosis is generally considered a silent disease and many patients are asymptomatic, certain symptoms and characteristic physical changes may alert the physician to the possible presence of osteoporosis. A patient with osteoporosis may present with acute lower back pain indicative of a fragility fracture of the vertebrae caused by activities of daily living (e.g., raising a window or making a bed). The patient may mention a history of wrist (Colles') fracture from minor trauma. These osteoporotic fractures are notable because they occur with a minor traumatic event.

The patient with multiple vertebral fractures over time, due to osteoporosis, may experience chronic back pain, loss of height, and kyphosis. Continued vertebral fractures can result in the ribcage tilting downward toward the hips, leading to a forward curvature of the upper spine, called kyphosis. This resulting deformity is also known as the dowager's hump. As this shift in posture continues, there is a compensatory anterior shift of the lower spine, called lordosis, resulting in protrusion of the abdomen. These characteristic changes in physical appearance can lead to a presumptive clinical diagnosis of osteoporosis (see figure below).

Progression of Osteoporotic Deformity

From Sinaki and Mikkelsen [1984]; with permission.
Click on image for animation/larger version.

Laboratory Tests and Biochemical Markers

Once the medical history and physical examination have been completed, diagnostic studies should be performed in order to confirm the presence, extent, and severity of low bone mass.

Routine Laboratory Tests

Laboratory tests are used to assist in the diagnosis of disease and to provide physiologic information about body processes. The following information discusses laboratory tests used in evaluating metabolic bone disease.

Most routine laboratory test results fall within the normal range for patients with osteoporosis. If routine tests are abnormal in a patient with low bone mass, special attention should be given to eliminating causes of secondary osteoporosis. Serum calcium, phosphorus, and alkaline phosphatase tests results are usually normal in primary osteoporosis. Urine calcium level is an inexpensive assay that is useful in detecting hypercalciuria (excess urine calcium loss) but is not usually useful in diagnosing primary osteoporosis.

Serum calcium and serum phosphorus levels

Routine evaluations of serum calcium and phosphorus levels are determined by automated spectrophotometry on blood samples. In most clinical laboratories, normal serum calcium levels range from 8.5-10.5 mg/dL.

The term phosphate is used interchangeably with phosphorus. Phosphate is phosphorus with oxygen molecules attached. Phosphate, the form of phosphorus that circulates in the blood, is transported across plasma membranes and filtered by the kidney. In most clinical laboratories, normal serum phosphorus levels range from 3.0-4.5 mg/dL.

Urine calcium levels

A 24-hour urine calcium specimen reflects calcium homeostasis. Normal values of 24-hour urine calcium are 100-250 mg, based on an average calcium intake of 600-800 mg/day.

An alternative test is a spot urine for calcium creatinine ratio in which a single urine sample (rather than a 24-hour urine specimen) is collected under fasting conditions. Hypercalciuria (excess calcium loss in urine, possibly seen with hyperparathyroidism) may contribute to excess skeletal loss of calcium. Conversely, a urine calcium level of <50 mg for 24 hours may indicate vitamin D malnutrition or malabsorption and inadequate calcium intake. Urine calcium levels vary considerably, even within an individual, which reduces the use of urine calcium levels for definitive diagnosis.

Parathyroid hormone levels

Parathyroid hormone (PTH) is the principal regulatory hormone for calcium in the body. By acting directly on the bone and the kidney, PTH can cause an increase in blood levels of calcium. Indirectly, PTH increases intestinal calcium absorption by activating vitamin D in the kidneys.

PTH is a polypeptide hormone that circulates in the body in several forms, including intact PTH and fragments of PTH. The fragments that exist in the circulation are the carboxy-terminal, or C-terminal, fragments of PTH and the nitrogen-terminal, or N-terminal, fragments of PTH. Intact PTH hormone possesses hormone activity. Older assays of PTH were less precise because they measured both intact PTH and PTH fragments. A two-site assay that more precisely measures intact PTH and excludes inactive fragments is available. PTH measurements should be interpreted together with simultaneous serum calcium levels for more accurate diagnoses of suspected parathyroid disorders.

PTH assays are important in the differential diagnosis of disorders of calcium metabolism, both hypercalcemia and hypocalcemia. The parathyroid glands can malfunction causing hyperparathyroidism or hypoparathyroidism. It is important to distinguish osteoporosis from hyperparathyroid disorders because the latter are generally best managed surgically. The following table shows the relationship between PTH values and these diseases.

Differential Diagnosis of Disorders of Calcium Metabolism

Disease	Serum PTH Level
Primary osteoporosis	Normal
Hyperparathyroidism	Elevated
Hypoparathyroidism	Depressed
Pseudohypoparathyroidism	Elevated

Click here for a review of the biochemistry of Biochemical Markers of Bone Turnover.

Biochemical Markers of Bone Turnover

Certain laboratory tests, called biochemical markers of bone turnover, have been developed to look at bone metabolism. Although greater emphasis may be placed on these laboratory tests in the future, the individual physician is not likely to routinely use these tests now. Bone turnover markers provide potentially useful information but cannot be used alone to diagnose osteoporosis, to determine the severity of the disease, or to select a specific therapy. They are widely used in clinical research to look at population responses to bone-active therapeutic agents. The following table provides a summary of the biochemical markers of bone turnover.

Biochemical Markers of Bone Turnover

Resorption Markers

From collagen degradation         Urine hydroxyproline     Urine total pyridinoline     Urine total deoxypyridinoline     Urine collagen type I cross-linked N-telopeptide; serum test also available     Urine collagen type I cross-linked C-telopeptide (Crosslaps)

Bone Formation Markers

From osteoblast secretion         Serum osteocalcin     Serum total alkaline phosphatase     Serum bone-specific alkaline phosphatase     From collagen formation         Serum procollagen I carboxyterminal propeptide (PICP)     Serum procollagen type I N-terminal propeptide (PINP)

Biochemical markers of bone turnover cannot be used to diagnose osteoporosis because markers cannot accurately distinguish between normal or low bone mass. The patient may have osteoporosis due to inadequate attainment of peak bone mass, rather than an increase in bone turnover. The patient may have had a prior period of high turnover rate and relatively rapid bone loss that no longer persists. An elevated bone turnover rate occurs in several different diseases and conditions (e.g., hypercortisolism, hyperparathyroidism, and long-term immobilization) or as a result of some types of drug therapy (e.g., anticonvulsants, glucocorticoids, excess thyroid hormone, gonadotropin-releasing hormone agonists, and heparin).

Thus, bone markers cannot be used to diagnose osteoporosis and distinguish between other diseases or conditions that may elevate the rate of bone turnover. Although biochemical markers are useful in osteoporosis research and in population studies, the use of such markers in individual patients remains controversial. Due to various factors, a single determinant of a biochemical marker may underestimate or overestimate the mean rate of bone turnover in an individual and may, or may not, provide clinically meaningful information. Over-reliance on markers could lead to misidentification of individuals. Additionally, variations in the biochemical marker measurements require caution and clinical judgment when making therapeutic decisions [Bikle, 1997].

Evidence exists to support the clinical use of biochemical markers of bone turnover as adjuncts to BMD testing in at least three major categories:

 
• as an independent factor for fracture risk
• as an aid to clinical decision making
• as an aid to assessing a patient’s biologic response to pharmacologic therapy.

Considerations in the use of biochemical markers of bone turnover

The use of biochemical markers in the diagnosis and management of osteoporosis is evolving, and the markers discussed may be supplemented or replaced by newer methods that offer various technical improvements or clinical advantages. Published medical literature recommends that biochemical markers of bone turnover be followed closely in postmenopausal women if the values are above the upper limit of the premenopausal reference range [Garnero and Delmas, 1996].

Some medical literature also recommends that if treatment for osteoporosis is initiated after documenting an elevated bone resorption marker, physicians should measure it again in 3-6 months to determine whether the patient has a biological response to therapy. The biochemical markers can be viewed as a surrogate indicator of a clinically meaningful response to therapy [Garnero and Delmas, 1996].

Biochemical markers of bone turnover should be used cautiously. The analytic precision of the assay technique and the natural biologic variability of the marker can affect the measurement. Variation within individuals over short (e.g., diurnal) and long time periods and variation between individuals affects the interpretation and establishment of normal reference ranges [Miller et al., 1999].

Selected factors can affect the results of biochemical marker assays:

• Changes in time of day (diurnal variation) can affect biochemical markers. Bone remodeling peaks at night. A first-void specimen may reflect peak activity. Correct timing of the specimen may be necessary for accurate results of a urine test. If a serum sample can measure the biochemical marker, accuracy may be improved, although urinary values reflect cumulative activity over a number of hours, rather than a single point in time [Bikle, 1997].
• Seasonal differences based on variation in vitamin D levels can affect biochemical markers [Miller et al., 1999].
• Urine measurements should be normalized to creatinine to adjust for muscle mass and renal function. Creatinine measurements are affected by analytical and biologic variation inherent in the measurement.
• Considerable individual variation in many markers occurs on a daily basis, which cloud the clinical use of such tests. Within-subject variability of urinary markers of bone resorption has been estimated to be about 10-29%, whereas the variability of serum markers of bone formation (specifically, bone-specific alkaline phosphatase, osteocalcin, and PICP) has been estimated to be about 12-17% [Garnero et al., 1994; Miller et al., 1999].
• Several biochemical markers are cleared through the kidney. Chronic renal failure causes accumulation of the markers and may produce a test result that falsely indicates increased level of bone activity [Bikle, 1997].
• Most clinical studies have assessed selected biochemical markers, making it difficult to compare the efficacy of one marker with the others [Bikle, 1997].

In summary, although biochemical markers have been useful in osteoporosis research and in population studies, their application in individual patients remains somewhat controversial. Because of the factors discussed, a single determinant of a biochemical marker may underestimate or overestimate the mean rate of bone turnover in a given individual and may or may not be associated with clinically meaningful effects. Increased use of markers could influence the cost of management of osteoporosis. Over reliance on markers could misidentify individuals. Additionally, because biochemical marker measurements vary, caution and clinical judgment should be used when making therapeutic decisions [Bikle, 1997].

Radiologic Techniques to Assess Bone

If a patient is suspected of having an osteoporotic fracture, standard X-rays of the affected part of the skeleton should be taken. Conversely, if no such fracture is felt to be present, standard X-rays are not needed. Standard X-rays are not generally diagnostic of osteoporosis unless a typical fracture exists. For several reasons, standard X-rays do not offer sufficient precision to assess low bone mass. Changes in bone mass are detectable on X-ray only when 30-50% of bone has already been lost [WHO Study Group, 1994]. This is partly because cortical bone is the type of bone most clearly seen on the standard X-ray, and extensive loss of trabecular bone can occur before such loss can be appreciated visually.

Although X-rays are useful for the diagnosis of osteoporotic fracture, including assessment of degree of vertebral deformity, they are not useful for the diagnosis of low bone mass.

Vertebral fractures can take several forms based on degree of deformity of the vertebrae, including biconcavity, wedging, and compression (crush) fractures. Vertebral fractures that are demonstrated on X-ray are called morphometric fractures, which typically include both symptomatic and asymptomatic fractures. In most cases the patient will bring the symptomatic vertebral fracture (clinical fracture) to the physician’s attention and then an X-ray will confirm it. Monitoring a patient over time with serial X-rays, as performed in clinical studies, will allow the identification of morphometric vertebral fractures and, therefore, will generally include both symptomatic and asymptomatic fractures.

Normal vertebral dimensions vary from vertebra to vertebra within individuals as well as across the sexes and across ethnic groups. Assessing change in vertebral shape requires monitoring the vertebrae by X-ray over a period of time and comparing, through careful measurement, the changes in dimensions to baseline. If a change occurs, the individual may have experienced a fracture.

Measurement of BMD

Importance of BMD

Much effort has been expended identifying individuals with or at risk for osteoporosis in an efficient and cost-effective way, to appropriately target intervention. Although certain risk factors are associated with low bone mass, being able to accurately predict who has osteoporosis by risk factors alone has not been as successful as hoped.

Obtaining a bone mass measurement (BMD testing) is the most accurate way to assess fracture risk [Kleerekoper, 1998; Levis and Altman, 1998]. BMD can also be used to help confirm the diagnosis of osteoporosis in patients with preexisting fractures [Levis and Altman, 1998].

T-Score and Z-Score

A patient’s BMD, measured by densitometry and expressed in g/cm², is compared with a "normal value." The normal value is the mean BMD of sex-matched young adults at their peak bone mass, sometimes referred to as the "young adult mean." When compared with the normal value, a patient’s BMD can be expressed in terms of the number of standard deviations (SD) from the normal value. A convenient way to express this is a T-score. The Z-score normalizes a patient’s BMD in a different way, by comparing the amount of bone loss with the expected loss for individuals of the same age and sex. The following graph depicts how T-scores compare with Z-scores in the evaluation of BMD.

Comparison of T-Score vs. Z-Score to Evaluate BMD

Click on image for larger version.

A typical example of a normal lumbar spine BMD value determined by dual-energy X-ray absorptiometry (DXA) for a 55-year-old woman is 0.95 g/cm². Because the particular BMD value obtained for a given patient may vary by as much as 10-12% when using equipment produced by different manufacturers, the BMD result for this patient could range from 0.84-1.06 g/cm².

A strong association exists between decreased BMD and increased fracture risk, for both vertebral and nonvertebral sites (e.g., hip and wrist) [Melton et al., 1993]. Fracture risk is inversely proportional to BMD. Based on population studies, for each 1 SD below the young adult peak mean bone mass, the risk of fracture increases 1.5-3-fold. Although BMD measurements at the spine or peripheral sites (i.e., wrist, finger, or heel) are useful in predicting hip fracture, measurement of hip BMD has greater predictive value in determining risk of future hip fracture than measurement of BMD at other skeletal sites. In one study, for every SD decrease below the young adult peak mean BMD at the femoral neck, there was a 2.6-fold increase in the risk of hip fracture after adjustment for age.

Interpreting a T-score

A patient’s BMD, measured by densitometry, is compared with the mean BMD of sex-matched young adults at their peak bone mass, sometimes referred to as the "young adult mean." When compared with the young adult mean, a patient's BMD can be expressed in terms of the number of SDs from the normal value. The T-score has been developed as a useful clinical tool to express the difference in SDs above or below the young adult mean peak BMD [Bracker and Watts, 1998]. The T-score is used to assess the patient for osteopenia or osteoporosis. The tables below show T-Score Interpretation and T-Score Interpretation Based on the WHO Criteria.

T-Score Interpretation

T-Score	Interpretation
Above -1 SD	Normal bone mass
-1 to -2 SD	Low bone mass (osteopenia)
Below -2 SD	Osteoporosis

Data from Nordin [1994].

T-Score Interpretation Based on the WHO Criteria

T-Score	Interpretation
Above -1 SD	Normal bone mass
-1 to -2.5 SD	Low bone mass (osteopenia)
Below -2.5 SD	Osteoporosis

From Bracker and Watts [1998]; with permission.

A T-score of -1 is one SD below young adult peak mean and represents an approximate 12% decrease in bone mass, regardless of the technique used or the site measured [Melton et al., 1993].

Definitions for Understanding BMD Measurements

Central BMD measures the axial and appendicular skeleton, which are the most commonly measured sites. The axial skeleton consists of the bone and cartilage in the head, neck, and trunk, whereas the appendicular skeleton consists of the shoulder blade and collarbone, the upper and lower limbs, and the pelvis. Peripheral BMD measures the most peripheral areas of the appendicular skeleton — namely, the forearm, phalanges, and calcaneus or os calcis.

Either central or peripheral BMD measurements can be useful in making a clinical decision regarding intervention for the prevention or treatment of osteoporosis. Both the American Association of Clinical Endocrinologists (AACE) [AACE and the American College of Endocrinology, 1996] and the National Osteoporosis Foundation (NOF) [1998] concur that BMD measurements at any peripheral or central site have value in fracture risk assessment.

The following figure shows the locations of the posterior and anterior sides of the vertebra, as well as the vertebral body and posterior processes.

Anatomy of a Vertebra

Adapted from Shier et al. [1996]; with permission.
Click on image for larger version.

A posterior-anterior (PA) lumbar spine BMD measurement is obtained when the radiation source is located on the posterior side of the body, directing radiation through the body from the posterior side toward the anterior side, and the detector is located anterior to the body (see figure below).

Posterior-Anterior Lumbar Spine Measurement

Click on image for larger version.

A lateral lumbar spine BMD measurement is obtained when the radiation source is located on one side of the body and the detector is located on the other side of the body. A lateral view of the lumbar spine is shown in the figure below.

Click on image for larger version.

As shown in the following figure, the PA spine BMD measurement includes the posterior processes of the vertebrae and the vertebral body. Degenerative changes in the spine, such as bony growths called osteophytes, may inaccurately elevate the BMD value obtained [Levis and Altman, 1998].

Directions of Radiation on Vertebra Undergoing BMD Measurement

Adapted from Shier et al. [1996]; with permission.
Click on image for larger version.

The lateral spine BMD measurement isolates the vertebral body and measures the trabecular and cortical bone in the vertebral body, but excludes the cortical bone in the posterior processes of the vertebra. This provides a better measurement of the area of interest, the trabecular bone, which is prone to fracture when bone density in the spine decreases.

Relationship of BMD to Fracture Risk

A strong association exists between decreasing BMD and increasing risk of fracture [Levis and Altman, 1998]. Numerous prospective studies have shown that measurement of BMD, using various BMD technologies at various skeletal sites (central or peripheral), accurately and precisely predicts the risk of future fracture [Black et al., 1992; Greenspan et al., 1997; Melton et al., 1993].

BMD is a continuous measure of fracture risk, just as blood pressure readings are a continuous measure of stroke risk. There is no absolute threshold value for either measurement. Just as there is no blood pressure reading above which a stroke will certainly occur or below which a stroke will certainly not occur; there is no T-score below which a fracture will certainly occur and above which a fracture will certainly not occur. Therefore, to make a therapeutic decision in clinical practice, a T-score should be used with the patient's medical history and clinical assessment, including information on additional risk factors for osteoporosis, concomitant diseases, and independent risk factors for fracture (e.g., risk of falling) [Kleerekoper, 1998].

There are some important considerations in using BMD measurements.

• Although a universally accepted optimal site of BMD measurement for the prediction of future fracture has not been established, there is evidence to suggest that BMD measurement at the site of interest usually provides the best prediction of fracture at that site. Measurement at the hip is the best predictor of hip fracture [Levis and Altman, 1998].
• Fracture risk varies at different anatomic sites. Measurement at each anatomic site leads to distinct fracture risk predictions — e.g., the predictability of vertebral fracture risk from vertebral BMD is different than the predictability of femoral neck fracture from femoral neck BMD.
• Discordant BMD results between measurements at different sites in the same person are common [Baran at el., 1997]. Many testing centers routinely measure BMD at both the spine and hip [Bracker and Watts, 1998; Kleerekoper, 1998]. A T-score at one site might be osteoporotic, whereas a T-score at another site may show normal BMD or osteopenia.
• Obtaining a BMD measurement at any site is useful in predicting future fracture risk and more useful than no BMD measurement [Levis and Altman, 1998].


Based on population studies, for each 1 SD below the young adult peak mean bone mass, there is about 1.5-3-fold increased risk of osteoporotic fracture at different skeletal sites [Melton et al., 1993]. The relationship of BMD and the site-specific prediction of fracture risk from BMD at that skeletal site, after adjustment for age, from a study of 304 women in Rochester, Minnesota, over 10 years, is as follows [Melton et al., 1993]:
• Each 1 SD decrease in lumbar spine BMD was associated with a 1.9-fold (190%) increase in the risk of a vertebral fracture.
• Each 1 SD decrease at the femoral neck or trochanter was associated with a 2.3-2.4-fold (230-240%) increase in the risk of hip fracture.
• Each 1 SD decrease in forearm BMD was associated with a 2.7-fold (270%) increase in the risk of a forearm fracture.

Measurement of hip BMD has greater predictive value in determining risk of future hip fracture than measurement of BMD at other skeletal sites, although BMD measurements at the spine or peripheral sites (i.e., wrist, finger, heel) are useful in predicting fracture at remote sites as shown below.

The following relationships of BMD and the site-specific prediction of fracture risk from the BMD value at a remote site are as follows [NOF, 1998]:

• Each 1 SD decrease in lumbar spine BMD was associated with a 1.3-fold (130%) increase in the risk of hip fracture.
• Each 1 SD decrease in lumbar spine BMD was associated with a 1.6-fold (160%) increase in the risk of wrist fracture.
• Each 1 SD decrease in forearm BMD was associated with a 1.6-fold (160%) increase in the risk of hip fracture.
• Each 1 SD decrease in forearm BMD was associated with a 1.6-fold (160%) increase in the risk of wrist fracture.
• Each 1 SD decrease in hip BMD was associated with a 1.9-fold (190%) increase in the risk of lumbar spine fracture.
• Each 1 SD decrease in hip BMD was associated with a 1.6-fold (160%) increase in the risk of wrist fracture.

Accuracy and Precision of BMD in Determining Actual Bone Mass

Several methods of determining measurement accuracy of the BMD technologies have been established and, in general, involve scanning an appropriate spine or femur segment in a cadaver and subsequently weighing the scanned bone after defatting and ashing that bone segment. Reports of the accuracy of BMD measurements through a comparison to the ash weight have been published for various commercially available instruments [Wahner et al., 1994]. Manufacturers routinely report statistics on the accuracy and precision of their instruments.

• Accuracy refers to the ability of the instrument to measure the same bone mineral content of a bone as measured by another method, such as ashing of the bone.
• Precision refers to the ability of the instrument to report the same results in repeated measurements [Wahner et al., 1994]. Although short-term precision error (obtained by repeating the test multiple times over a short time period) may be small (about 1%), estimates for the precision error in clinical practice over a long time period may be slightly larger (1.3-1.8%) [Wahner et al., 1994].

Instrument precision and an individual’s change in BMD

Within the limits of the precision of the instrument, repeated measurements of bone density provide information on the rate of bone loss and the response to pharmacologic therapy [Levis and Altman, 1998].

With a precision error of 1%, a change of 1% in a BMD measurement from baseline to a later time point may not reflect a real change in BMD. If an instrument with a precision error of 1% is used, the patient must have a measured BMD change of 1.8% to be 90% confident that the change is real and not due to measurement error. To be 95% confident that the change in BMD is real, the patient must have a BMD change of 2.8% [Wahner et al., 1994]. The more often a BMD measurement is obtained, the greater the precision.

Evolution of BMD Technology

The noninvasive approaches to BMD measurement have been evolving as scientific research resulted in new discoveries and increased knowledge. The following information should be regarded as a snapshot in time and the instrumentation discussed below may be replaced in the future by newer methods.

Technical advancements in the instrumentation are sought to:

• improve the accuracy and precision of the bone mass measurement
• decrease patient exposure to radiation
• improve patient acceptability
• enhance portability
• increase the range of skeletal sites that can be measured
• reduce cost so that the instruments can be more widely available for diagnosis.

The following lists the technologies in the approximate order they were developed for measurement of bone mass:

• Radiographic absorptiometry (RA)
• Quantitative computed tomography (QCT)
• Single-photon absorptiometry (SPA)
• Dual-photon absorptiometry (DPA)
• Single X-ray absorptiometry (SXA)
• Dual X-ray absorptiometry (DXA)
• Quantitative ultrasound (QUS).

Basic Principles of BMD

The principles of bone mass measurement are based on the principle that bone attenuates or absorbs ionizing radiation. Therefore, the more bone present, the more radiation the bone absorbs. The unabsorbed radiation passes through the bone and is measured in a radiation detector [Levis and Altman, 1998]. This is the basic principle underlying most of the noninvasive radiologic bone mass measurement techniques, such as:

• SPA and DPA
• SXA and DXA
• QCT.

A radiation-free technique for BMD measurement that detects the transmission of high-frequency sound waves through bone is QUS. Because the velocity of sound is higher in healthy bone, QUS can measure bone mass and give some information about bone microarchitecture [Levis and Altman, 1998]. Future research will explore the use of QUS to determine the microarchitecture of bone.

The development of new techniques that safely yield precise measurements of bone mass at specific skeletal sites has contributed to increased testing of patients for osteoporosis.

A limitation of photon absorptiometry and X-ray absorptiometry is that these measurements reflect a two-dimensional estimation of a three-dimensional object. The values obtained in terms of bone mass are expressed in grams per area in centimeters squared (g/cm²) rather than bone mass in grams per volume in centimeters cubed (g/cm³). Therefore, these techniques do not measure true volumetric bone density, which is mass per volume (g/cm³). In contrast, QUS and QCT express results in terms of mass per volume (g/cm³) [Levis and Altman, 1998]. Nevertheless, bone mineral density is the accepted term for the results of all bone assessment techniques.

Bone density, however measured, is related to normal values specific for the patient's sex and anatomic site and the type of instrument used. Taken into account with the age and medical history of the patient, bone density measurement can be used to determine a patient's risk of future fractures [AACE and the American College of Endocrinology, 1996].

Early methods used to estimate bone density include conventional radiography, radiogammetry, radiographic photodensity, SPA, and DPA. Each of these methods has largely been replaced in clinical practice by faster, more accurate, more precise, or less expensive methods.

Single energy X-ray absorptiometry

SXA uses an X-ray tube to produce a single photon beam. To account for soft-tissue absorption, the body part is immersed in a water bath that stimulates a uniform soft-tissue thickness. Compared with SPA, SXA results in higher resolution, shorter scanning times (10-15 minutes), and reduced doses of radiation to the patient. SXA is used to image distal skeletal sites, such as the calcaneus. The following figures show the area of the heel measured in SXA imaging, an SXA image of the heel, and a typical SXA report.

SXA has excellent longitudinal precision of 1-2%. Accuracy is 2-5%. SXA has largely replaced SPA. Bone mineral content is measured in terms of grams per centimeter squared (g/cm²) [AACE and the American College of Endocrinology, 1996; Jergas and Genant, 1993].

Area of Heel Measured in SXA Imaging

Click on image for larger version.

SXA Image of the Heel

Click on image for larger version.

Typical SXA Report

Click on image for larger version.

Dual energy X-ray absorptiometry

The most recent advance in absorptiometry has been DXA. Instead of using an X-ray tube that produces a single photon beam as in SXA, DXA uses an X-ray tube that produces two photon energy beams. This results in higher precision (1-3%), shorter scanning times (<5 minutes), reduced doses of radiation to the patient (much less than the dose of a standard chest X-ray), and an accuracy of 1-10%. DXA measurements can be performed at central sites, such as the spine and hip, or at peripheral sites, such as the forearm or wrist. Bone mineral content is measured in terms of grams per centimeter squared (g/cm²) [AACE and the American College of Endocrinology, 1996; Jergas and Genant, 1993].

DXA has become the method of choice, or gold standard, for measuring bone mass in the spine and femoral neck and has largely replaced DPA [AACE and the American College of Endocrinology, 1996; Levis and Altman, 1998]. Also, DXA can now perform scans with radiologic quality that gives increased information about the trabecular content in lateral projections of the spine. DXA also provides total body scans for bone mass and body composition. Some DXA machines can provide both PA lumbar spine and lateral lumbar spine BMD measurements. However, some DXA machines can only provide PA lumbar spine measurements, which may result in inaccurately elevated BMD values if certain degenerative changes in the spine are present [Levis and Altman, 1998].

The major disadvantages of central DXA are its high cost, need for dedicated space, and lack of portability. Also, trained and qualified technologists are required for optimum test performance [AACE and the American College of Endocrinology, 1996; Kleerekoper, 1998].

Importantly, central DXA is used for comparison or follow-up studies to monitor response to therapy [AACE and the American College of Endocrinology, 1996]. The following figures show a DXA image of the hip, a DXA image of the lumbar spine, and a typical DXA report.

DXA Image of the Hip

Click on image for larger version.

DXA Image of the Lumbar Spine

Click on image for larger version.

Typical DXA Report

Click on image for larger version.

Peripheral DXA

Peripheral DXA (pDXA) measures peripheral sites only. It has several advantages: ease of use, smaller size, lower cost, and shorter measurement time. However, peripheral DXA is not recommended for monitoring response to therapy at this time [Bracker and Watts, 1998]. The following figures show a sample pDXA image and a typical pDXA report.

Sample pDXA Image

Click on image for larger version.

Typical pDXA Report

Click on image for larger version.

Quantitative Computed Tomography

QCT provides an image of a thin transverse slice through the body. QCT is a technique available that measures true volumetric bone density (e.g., a three-dimensional measurement expressed in g/cm³) [Baran et al., 1997; Jergas and Genant, 1993].

The images in QCT are derived from measurements of tissue attenuation. The amount of energy that is attenuated or absorbed can be measured. The attenuation is dependent on tissue density and composition, allowing for distinct measurement of both trabecular and cortical bone density of several sites in the body. The tissue attenuation values are compared with an external, bone mineral reference phantom. QCT of the spine uses a conventional CT scanner with a calibration phantom and special software to measure vertebral bone mass. A lateral QCT localizes the midplane of two to four lumbar vertebral bodies and quantitative readings are then obtained. QCT determination of density in two vertebrae is then compared to known density readings of external phantoms. The measurements of the vertebrae are then averaged and used to calculate the density of trabecular bone in the vertebrae [Jergas and Genant, 1993].

QCT is available in a single-energy mode with a radiation dose less than or equal to standard chest X-ray and a dual-energy mode, which has a higher radiation dose. The radiation of single-energy QCT is typically equivalent to a transcontinental airline trip. Although dual-energy QCT improves accuracy, it is only recommended for research studies because of its higher radiation dose [Baran at el., 1997; Jergas and Genant, 1993].

Unlike PA lumbar spine BMD measurements obtained by DXA, lumbar spine BMD measurements obtained by QCT are not influenced by degenerative bony changes. Total scanning time for QCT is several minutes. Disadvantages of QCT include care in calibration and positioning, expensive cost, and higher dose of radiation as compared to other techniques [AACE and the American College of Endocrinology, 1996; Jergas and Genant, 1993; Levis and Altman, 1998]. The following figures show a QCT image of a lumbar vertebra and a typical QCT report.

QCT Image of a Lumbar Vertebra

Click on image for larger version.

Typical QCT Report

Click on image for larger version.

Peripheral QCT

QCT is available in smaller-sized machines, which are portable and capable of peripheral measurements. QCT at peripheral sites (pQCT) can distinguish between cortical and trabecular bone and measures forearm BMD [Jergas and Genant, 1993]. The following figure shows a typical pQCT report.

Typical pQCT Report

Click on image for larger version.

Radiographic absorptiometry

RA provides a practical and rapid way of determining bone mineral density of the middle phalanges. X-rays are taken of the patient’s hand using standard X-ray equipment. As shown in the figure below, a reference wedge of aluminum alloy is placed next to the patient’s hand and is, therefore, filmed at the same time. Sections of the X-ray of the patient’s middle phalanges of the fingers are then optically scanned. A computer processes the data. The patient's BMD is determined by comparing the optical density of the bone with that of the calibrated aluminum alloy wedge. Corrections for factors such as soft tissue absorption and positioning of the patient are incorporated into the calculations [Kleerekoper, 1998; Levis and Altman, 1998].

The Technique of Radiographic Absorptiometry

Adapted from Digital Imagery [2000] PhotoDisc Inc.; with permission.
Click on image for larger version.

Although radiographic absorptiometry measures BMD at a peripheral site (i.e., finger) that is not clinically relevant to osteoporosis, the BMD of the phalanx portion of the finger does provide good correlation with central BMD. This is probably because the phalanx is made up of roughly equal proportions of trabecular (40%) and cortical bone (60%). One study found that RA predicted low bone mass with a sensitivity of 90% for the lumbar spine and 82% for the femoral neck. This technology lacks the ability to determine BMD at other anatomic sites, such as lumbar spine and hip. Despite its advantages of using readily available technology and needing minimal training to perform the test, radiographic absorptiometry has had limited use.

Quantitative ultrasound

QUS is a radiation-free technique to measure bone mass. QUS detects the transmission of high-frequency sound waves through or across bone. QUS measures two characteristics of sound in bone:

• Broadband ultrasound attenuation (BUA)
• Velocity of sound (VOS) or speed of sound (SOS).

BUA is used to determine bone density and to gain some information about bone structure. It is based on the principle that the more complex the structure, the greater the attenuation of the ultrasound (e.g., the more the sound is absorbed) [Levis and Altman, 1998]. The following figures show BUA in healthy bone and BUA in osteoporotic bone, respectively.

Broadband Ultrasound Attenuation in Healthy Bone

Click on image for larger version.

Broadband Ultrasound Attenuation in Osteoporotic Bone

Click on image for larger version.

VOS evaluates the bone density and elasticity by measuring how fast sound will go through the bone. The following figure shows the movement of sound waves through the calcaneus. The greater the connectivity of the trabeculae, the faster the sound travels. Another way of saying this is that attenuation of ultrasound and the velocity of sound are higher in healthy bone than in osteoporotic bone. Because the velocity of sound in bone is related to the elasticity, density, and separation of the trabeculae, ultrasound may also provide structural information about the bone (e.g., bone microarchitecture) [Levis and Altman, 1998].

Sending and Receiving of Ultrasound Signals through the Calcaneus

Click on image for larger version.

Some QUS systems report the ratio of BUA to SOS as the quantitative ultrasound index (QUI) or stiffness index (SI). Both QUI and SI are considered BMD equivalents. Measurements using QUS are usually carried out at the calcaneus. The following figure shows a typical QUS report.

Typical QUS Report

Click on image for larger version.

The BUA and SOS may also be reported. Although report terminology for devices may differ, the ability of all QUS devices to predict fracture risk is similar.

Studies have shown good correlation between QUS results and DXA results of the lumbar spine [Bauer et al., 1995] and the femoral neck [Bauer et al., 1995; Hans et al., 1996]. Cited advantages of QUS include rapid measurements, portability, radiation-free assessment, and minimal cost. QUS may also provide information on bone structure and strength independent of BMD, as measured by bone densitometry methods.

QUS allows good discrimination between patients with osteoporosis and normal patients. Because of the higher variability of this technique, further study to distinguish those patients with low bone mass and false negative results is needed [Greenspan et al., 1997]. Furthermore, statistical correlations with DXA measurements and appropriate T-score cutoffs for fracture prediction need to be better defined through prospective studies.

QUS is considered an alternative to DXA for the initial screening and evaluation of fracture risk. However, because of lower reproducibility and limited data