 |
|
|
Bone mass loss and osteoporosis are associated with various conditions, such as end-stage renal disease (ESRD), and treatments, such as prolonged steroid therapy. Bone densitometry is used to measure bone mass density to determine the degree of osteoporosis and to estimate fracture risk. Bone densitometers measure the radiation absorption by the skeleton to determine bone mass of the peripheral, axial, and total skeleton. Common techniques include single-photon absorptiometry (SPA) of the forearm and heel, dual-photon (DPA) and dual-energy x-ray absorptiometry (DXA) of the spine and hip, quantitative computed tomography (QCT) of the spine or forearm, and radiographic absorptiometry (RA) of the hand. Part I of this report addresses important technical considerations of bone densitometers, including radiation dose, site selection, and accuracy and precision, as well as cost and charges. Part II evaluates the clinical utility of bone densitometry in the management of patients with ESRD. End-stage renal disease affected more than 242,000 Americans in 1992, and each year 10,000 to 20,000 new cases are diagnosed. Although the survival rate of ESRD patients has improved, metabolic bone diseases that fall under the generic term "renal osteodystrophy" represent abnormal development of bone and major long-term complications. Issues addressed are the type and extent of bone loss associated with ESRD and whether these patients have an increased risk for fracture. The other assessments in this series address the clinical utility of bone densitometry for patients with asymptomatic primary hyperparathyroidism, steroid-dependent patients, estrogen-deficient women, and patients with vertebral abnormalities.
The Center for Health Care Technology (CHCT) evaluates the risks, benefits, and clinical effectiveness of new or unestablished medical technologies. In most instances, assessments address technologies that are being reviewed for purposes of coverage by federally funded health programs.
CHCT's assessment process includes a comprehensive review of the medical literature and emphasizes broad and open participation from within and outside the Federal Government. A range of expert advice is obtained by widely publicizing the plans for conducting the assessment through publication of an announcement in the Federal Register and solicitation of input from Federal agencies, medical specialty societies, insurers, and manufacturers. The involvement of these experts helps ensure inclusion of the experienced and varying viewpoints needed to round out the data derived from individual scientific studies in the medical literature.
CHCT analyzed and synthesized data and information received from experts and the scientific literature. The results are reported in this assessment. Each assessment represents a detailed analysis of the risks, clinical effectiveness, and uses of new or unestablished medical technologies. If an assessment has been prepared to form the basis for a coverage decision by a federally financed health care program, it serves as the Public Health Service's recommendation to that program and is disseminated widely.
CHCT is one component of the Agency for Health Care Policy and Research (AHCPR), Public Health Service, Department of Health and Human Services.
Thomas V. Holohan, M.D., FACP Director Center for Health Care Technology
Clifton R. Gaus, Sc.D. Administrator Agency for Health Care Policy and Research
Questions regarding this assessment should be directed to:
Center for Health Care Technology
AHCPR
Willco Building, Suite 309
6000 Executive Boulevard
Rockville, MD 20852
Telephone: (301) 594-4023
Various medical conditions such as end-stage renal disease (ESRD) and treatments such as long-term steroid therapy and thyroid hormone replacement can result in bone mass loss and osteoporosis, which is often undetected until the person sustains a fracture. About 1.3 million fractures that occur annually in the United States in people over age 45 are ascribed to osteoporosis (1.5 million is the extrapolated value that accounts for the aging population).(1) Fractures exact an enormous toll on human disability and suffering as well as increase health care costs.
Bone densitometry is used to measure bone mass density to determine the degree of osteoporosis and fracture risk. Commonly used techniques are: single-photon absorptiometry (SPA) of the forearm and heel, dual-photon and dual-energy x-ray absorptiometry (DPA and DXA) of the spine and hip, quantitative computed tomography (QCT) of the spine or forearm, and radiographic absorptiometry (RA) of the hand.
Continuing advances in bone densitometry techniques prompted the Health Care Financing Administration (HCFA) to ask the Office of Health Technology Assessment (renamed the Center for Health Care Technology [CHCT]) to assess the effectiveness of SPA, DPA, DXA, QCT, and RA for five classes of patients. The OHTA previously assessed the general safety and effectiveness of SPA, DPA, and RA.(2 -4)
Whether use of bone densitometry to document bone loss effectively alters the medical management of patients to minimize the development of osteoporosis and fractures remains controversial. The debate continues over selection of the best measurement sites and the availability of effective treatment to prevent progression of osteoporosis if significant bone loss is detected. Some concern also remains about the accuracy and precision of these techniques. These uncertainties have led to a lack of specific recommendations and protocols for bone mass testing.
The purpose of this report is to assess which techniques, if any, are clinically useful in the medical management of patients with bone loss due to ESRD. Part I of the report describes each of these techniques; part II discusses their clinical utility in the management of ESRD patients. The other four assessments in this series will evaluate the clinical utility of bone densitometry for patients with asymptomatic primary hyperparathyroidism, steroid-dependent patients, estrogen-deficient women, and patients with vertebral abnormalities.
Since the 1960s, there has been a gradual development of noninvasive techniques to quantify bone mass density. Bone densitometers (absorptiometers) measure the radiation absorption by the skeleton to determine bone mass. The term "bone mass" indicates the amount of mineralized tissue in bone; the term "bone density" indicates the mass of bone defined either by length (grams per centimeter [g/g cm]), area (grams per square centimeter [g/gcm2]), or volume (grams per cubic centimeter [g/gcm3]).
The geometric distribution of bone tissue, together with its material properties, is critical in determining the structural strength and rigidity of the whole bone. Although a variety of factors may influence bone strength and the probability of fracture, bone mass measurements are currently considered to be the most valuable objective indicator of fracture risk.(5,6)
Bone densitometers can evaluate bone density of the peripheral, axial, and total skeleton. Their clinical use is based on the assumption that bone mass is an important determinant of fractures, and that bone mass measurements may help reduce the number of fractures by identifying high-risk patients, who can then receive effective prophylaxis. Because osteoporosis is generally considered preventable but not reversible, early detection of at-risk individuals is essential.(7, 8) Bone densitometry can also be used to document bone mass changes over time to monitor the course of a disease and response to therapy.
Factors that influence bone strength include the amount (mass), type (quality), and architecture (structure) of mineralized bone tissue. Adult bone tissue is primarily arranged either as dense cortical bone (also known as compact bone) or as a latticework of trabecular bone (also known as spongy or cancellous bone).
The combination of a smooth, dense, outer layer of compact cortical bone, together with spongy trabecular bone, provides both structural strength and an extended surface on which rapid changes in bone formation or resorption can respond to fluctuating metabolic needs. The metabolic responses of the skeleton mainly occur on trabecular bone surface.(9) Although the entire skeleton loses bone mass with aging, the distribution of bone loss is not uniform because of the different proportions of trabecular and cortical bone in the various parts of the skeleton.(10)
Osteopenia refers to any condition involving reduced bone mass. Osteoporosis, a form of osteopenia, is decreased bone mass with normal bone mineralization. Bone densitometry generally has been applied to patients with bone loss associated with primary osteoporosis, which mainly occurs in postmenopausal women and the elderly. Secondary osteoporosis, which occurs in less than 5 percent of those with osteoporosis, can be a consequence of such treatment as long-term steroids or such conditions as chronic renal failure, rheumatoid arthritis, and hyperparathyroidism.
Primary osteoporosis is commonly classified as postmenopausal osteoporosis (type I) or age-related osteoporosis (type II).(11) This classification was presented by Riggs and Melton,(12) but not all investigators agree with it, and alternative proposals have been developed.(13) Type I osteoporosis occurs in a subset of postmenopausal women between the ages of 51 and 75 years, affects mainly trabecular bone, and is largely responsible for vertebral crush fractures and Colles' fractures.(14)
Type II osteoporosis is found in a large proportion of women and men over the age of 70 and results from a protracted slow phase of age-related bone loss. This condition affects twice as many women as men and is manifested mainly by hip and vertebral fractures.(12,14)
Historically, osteoporosis was recognized clinically only after fractures occurred. It was considered a syndrome consisting of both substantial bone loss and fractures. (15,16) With the development of techniques that measured bone loss before signs of radiographic change or the occurrence of fractures, osteoporosis was defined as an age-related disorder characterized by decreased bone mass and by increased susceptibility to fractures.( 17)
How osteoporosis is defined is not a trivial issue. As a result of the various criteria used to define osteoporosis, estimates of its prevalence differ markedly. For example, if bone mass lower than 2 standard deviations (SD) below the mean of normal young women represents osteoporosis, then 45 percent of white women aged 50 years and older are osteoporotic.(18) A World Health Organization (WHO) study group recently defined osteoporosis as 2.5 SD below that mean, which identifies 30 percent of all postmenopausal women as osteoporotic, more than half of whom will have sustained a prior fracture.(19)
There is a growing tendency to define osteoporosis in terms of a continuum of bone density, with greatest fracture risk in those with lowest absolute density values, rather than in considering osteoporosis as a fracture/gnonfracture dichotomy. (20) However, so-called "fracture thresholds" have been approximated for various skeletal sites to diagnose osteoporosis based on a level of bone mass above which nontraumatic fractures rarely occur and below which fractures are more common (e.g., 90 percent of women with vertebral fractures have spinal bone density values less than .97 g/gcm2 measured by DPA).(21,22)
Other approaches include Nordin's proposal that osteoporosis be defined by identifying the normal range of bone density at specific anatomic sites in young adults and classifying bone density values less than 2 SD below the mean of this range as osteoporotic.(23) Although disagreement continues, the definition developed at the Consensus Development Conference,(24) held in Copenhagen in April 1993, is generally accepted. That conference defined osteoporosis as a disease characterized by low bone mass and microarchitectural deterioration of bone tissue, with a consequent increase in bone fragility and susceptibility to fracture.
Although osteoporosis affects all bones, the most common so-called osteoporotic-related fractures are those involving the spine (thoracic and lumbar vertebral bodies); femur (the neck and intertrochanteric regions); and wrist (distal radius).(1) In persons over the age of 50, the incidence of these fractures increases markedly. Fractures of the distal radius, common in the middle-aged and elderly, are caused by a fall on the outstretched hand. Approximately 172,000 osteoporotic fractures occurred in the distal radius during 1985.(25) The most severe osteoporotic fracture is that of the hip, a disabling and sometimes fatal event. The prevalence of hip fractures in the United States reaches 15 percent by age 80 and exceeds 25 percent by age 90.(1) Between 250,000 and 300,000 hip fractures occur each year in the United States and the incidence may be increasing.(26-28)
However, even among elderly persons with low bone mass, most falls do not result in a fracture.(29) Greenspan et al(30) found that, in most elderly fallers with hip bone density already below the so-called fracture threshold (>2 SD lower than average peak bone mass as measured by DXA), fall characteristics and body habitus were important risk factors for hip fracture. About 90 percent of hip fractures are the result of falls, but fewer than 1 percent of falls result in hip fractures.( 31)
Nevertheless, a hip fracture may often be a marker for declining health and impending death.(32) Most patients who suffer a hip fracture have severe concomitant illness. Browner et al(33) have found that diminished bone mass density is associated with increased all-cause mortality, even after excluding women with deaths resulting from fractures.
Although the true incidence of vertebral fracture is not known, more than 25 percent of all women over age 65 may have sustained a crush fracture, most of which are asymptomatic.(1) Fractures occurring at other sites, including the rib, pelvis, humerus, and proximal tibia also may be attributable to osteoporosis.(34)
Using a MEDLINE search of the National Library of Medicine data base, CHCT retrieved published articles in the English language from 1986 to 1994 addressing bone densitometry as well as its use in the diagnosis and treatment of osteoporosis related to ESRD. Text words used singularly or in combinations were osteoporosis, densitometry, absorptiometry, bone mass, bone density, measurement, bone site, precision, bone loss, biochemical markers, fracture, treatment, therapy, bone disease, end-stage renal disease, dialysis, renal osteodystrophy, QCT, DXA, RA, SPA, and DPA.
The CHCT also examined the bibliographies of reviewed articles to identify publications not otherwise captured. In addition, respondents to a notice published in the Federal Register forwarded or suggested many articles. The CHCT also solicited data from other agencies of the Public Health Service (PHS).
Draft copies of the assessment were reviewed by CHCT staff and PHS officials selected by the National Institutes of Health (NIH), the Food and Drug Administration (FDA), and the Centers for Disease Control and Prevention (CDC).
Bone densitometry techniques, previously used as research tools to elucidate metabolic bone diseases, are being used with increasing frequency in clinical practice. These techniques measure radiation absorption by skeletal calcium to determine bone mass density. Single-photon absorptiometry became available in the early 1960s. It measures both cortical and trabecular bone and is limited to measuring appendicular sites such as the wrist. Dual-photon absorptiometry emits photons of two distinct wavelengths and energy levels and can scan bones in two dimensions. It became commercially available around 1980 and is most often used to measure the lumbar spine or proximal femur.
Both SPA and DPA use radioisotopes as their photon source for imaging bones. Dual-energy x-ray absorptiometry, QCT, and RA use x-ray sources. Like SPA and DPA, DXA cannot separate cortical from trabecular bone. However, it can scan an area much more rapidly and is more precise than DPA, which it is quickly replacing.(35) Although widely available, DPA devices are no longer manufactured.(36)
The advantage of QCT is its ability to separate cortical from trabecular bone with great anatomic detail. On the other hand, it is a much more difficult scan to perform and emits 100 times or more the radiation of DXA. Finally, RA is used to measure bone density in the hand.
As yet, there is no agreement on which method or methods are most clinically effective for diagnosing and monitoring the individual patient or for identifying the high-risk patient for therapeutic intervention.
The FDA has allowed SPA, DPA, DXA, and QCT to be marketed under Section 510(k) of the Medical Device Amendments of 1976 to the Federal Food, Drug, and Cosmetic Act of 1938. For QCT, approval is solely for a software package that adapts computed tomography (CT) for bone density measurements. These devices may be marketed for measurement of bone mass and for monitoring an individual patient over time. The FDA regulates RA because it considers RA a medical device. Currently, there is no FDA approval for any bone density technique that claims to predict fracture risk.
| Equipment cost charges ($)/gsource | Scan charges ($) | Site/gscan time error (minutes> | Precision error (percent) | Accuracy error (percent) | Radiation exposure (millirem) | |
|---|---|---|---|---|---|---|
| SPA | 20,000-30,000/g750/4-6 mo | 50-150 | Radius, calcaneus; 5-15 | 2-5 | 3-8 | 2-5 |
| DPA | 30,000-65,000/g5,000-7,000/12-18 mo | 150-300 | Spine, hip; 20-40 | 2-5 | 3-10 | 5-10 |
| QCT | 5,000-15,000b | 150-400 | Spine; 10-30 | 2-6 | 5-15 | 100-1,000 |
| DXA | 60,000-100,000 | 150-300 | Spine, hip; 5-10 | 0.5-3 | 3-9 | < 5 |
| RA | Existing x-ray equipment | 90-160 | Hands; 5-10 | 2-4 | 5-10 | 10-100 |
a Adapted from Tables 3-7 in this assessment.
b To adapt an existing CT scanner, dedicated scanners are recommended for measurements of bone density.
Important differences in the characteristics and performance of these techniques include procedure charges, radiation exposure from the scan, the bone and bone site measured, and the accuracy and precision achieved. All these factors can affect the clinical utility of a given technique for managing specific patients.
Charges vary for different techniques as well as for the same technique used at different locations. Charges reported in the literature usually represent those of universities or research institutions. Smaller clinics and individual physicians not concerned with spine or hip measurements tend to install the less expensive SPA units. The more expensive DPA and now DXA units are usually installed in larger clinics, hospitals, and university settings or wherever spine or hip measurements are desired. DXA has the potential for lower costs, because DXA bone scans require much less time than QCT and DPA (which is being replaced), allowing for more patient measurements per unit of time.37 Also, the long life of the x-ray tube results in lower maintenance costs. Computed tomography units, which are widely available, can be adapted to perform QCT bone density measurements.
All bone densitometry techniques involve patient exposure to radiation. However, these techniques are generally considered to be safe, with relatively low radiation exposure to the patient. In longitudinal studies, cumulative dose becomes important. Single-photon absorptiometry, using an iodine (125I) radiation source, has the lowest dose per measurement (2-5 mrem). This dose is limited to the actual measurement sites, for example, the wrist and heel (calcaneus). Furthermore, there is no radiation-sensitive hemopoietic marrow in the calcaneus, as opposed to the vertebrae. Dual-photon absorptiometry, using gadolinium (153Gd) as the radiation source, usually delivers 5-10 mrem to the hip or spine with less than 1 mrem absorbed radiation dose at the gonads. Dual-energy x-ray absorptiometry doses are similar to those of SPA but involve exposure to the torso in the pelvic region. Quantitative CT delivers the greatest dose (100-1000 mrem; 100-300 mrem for newer scanners). Although about equal to a chest x-ray, the QCT dose is 20-200 times that of the other techniques, with the scattered gonadal dose also proportionally higher.5 This makes it more problematic for repeat measurements.
Although cost and radiation exposure are considerations that may influence the choice of a measurement site, the most important consideration is the ability to indicate fracture risk. Presently, there are insufficient longitudinal data to determine if any technique or skeletal site is better than others for quantifying bone mass or measuring bone loss rates.(38) Controversy continues regarding the best choice of a bone mass measurement site for determining the extent of osteoporosis and predicting fracture risk. Some studies conclude that, to predict fracture risk in a bone, the bone density of that specific bone should be measured, and that measurements made at other parts of the skeleton have less predictive value.(39) Conversely, others conclude that patients with vertebral fractures have generalized osteoporosis. Therefore, the measurement of spinal bone density has no predictive advantage over that of the forearm, and the risk of vertebral fracture can be predicted from appendicular sites as well as the spine (axial skeleton).(40- 42) This position is supported by the results of several recent prospective studies(6,43-46) that found no convincing evidence that measurements made at the spine and hip are superior to appendicular sites for predicting subsequent fractures at the spine and hip. Yet another study(47) concludes that low bone mass is generalized in some women and more regional in others and that those with the lowest bone mass measurements at any site are more likely to have extensive osteoporosis.
The ability to use a single-site measurement of bone density to evaluate general fracture risk at all sites has practical importance. It could eliminate the rationale for multiple-site measurements performed by some clinicians and would put less emphasis on a technique's ability to measure a particular site. Techniques not capable of bone mass measurements at the spine and/gor the hip would not be eliminated from consideration for that reason. For example, measurement of appendicular sites is considered easier, costs less, and can be done with most techniques. However, vertebral sites must be measured with DXA or QCT, and these measurements can be difficult to reproduce in patients with spinal deformities and aortic calcification, artifacts that may confound the bone mass measurements. This has led some to recommend the hip, forearm (radius), and heel as alternative sampling sites. (48)
Site selection is also an issue when bone densitometry is used to monitor response to therapy, because the cortical/gtrabecular composition of the measurement site must reflect the effects of particular therapeutic agents.(49) For example, Mazess et al(50) consider the spine the most appropriate site for bone loss monitoring, because it shows larger decreases in bone mass at the menopause than do appendicular sites and larger increases in bone mass after estrogen-replacement therapy. However, others suggest that such a distinction may be unnecessary and that measurement of bone density by SPA or QCT of the forearm could be used for clinical followup of bone changes. (51,52)
Measurements obtained by densitometers are compared with an independent standard measurement of bone mass, such as ashed bone sections or phantom standards calibrated against ashed bone sections.(53) The accuracy error is determined by how much the measurement varies from this accepted or "true" value. Precision error is the variability in the measurements occurring with repeated measurements of the same object.(54)
Concern has been expressed about the accuracy and precision of various techniques, particularly in the clinical setting. Single-photon absorptiometry has the greatest accuracy in vivo; DXA is considered the most precise, followed by SPA. The accuracy and precision of all the techniques can be reduced by improper patient positioning, changes in the strength of the radionuclide sources (SPA, DPA), improper calibration of the equipment, and weak quality assurance programs.
A technique's precision is critical for serial measurements that correctly document bone loss over time. A number of factors influence serial measurements, regardless of the technique chosen or site selected to determine bone mass. Attention to patient repositioning is necessary to assure that the same exact bone location is measured each time. If sequential scan locations are not matched, erroneous measurements may result, especially in regions of bone where bone mass is not uniform.
When the bone mass is expressed as an "area" density (SPA, DPA, DXA), the limb must have the same rotation and axis at each measurement so that the projected area will be the same. When the forearm, spine, hip, or heel are measured, variations in rotation or angulation of the axis will adversely affect measurements.( 5)
Changes in soft tissue composition can also have significant effects on bone mass measurements. A change in soft-tissue composition surrounding the bone to be measured changes the baseline reference values against which the bone absorbance values are compared. This is true equally for RA, SPA, DPA, and DXA. For example, changes in SPA measurements may occur with treatment that alters body fat, such as anabolic steroids, norethindrone, and estrogen.(55) An increase in radial density as determined by SPA may reflect a decrease in the subcutaneous forearm fat rather than an increase in bone mass. Furthermore, the fat content of red marrow in the spine and femur changes with age and can influence absorbance. As a result, errors in QCT measurements of these sites generally increase in older patients.(5)
Changes in the strength of radionuclide sources (SPA, DPA) can significantly affect the precision of serial measurements. Serial DPA measurements can be subject to errors of up to 6 percent due to age (diminished activity) of the radionuclide source.(3,56,57) The relatively low and constantly decreasing intensity of photon sources leads to problems with long-term stability of these instruments. Instrumental instability and problems with software changes have now been recognized. Yet, it is still unclear how well certain calibration standards are able to correct for changes in radiation source intensity.(58)
The ability of densitometers to measure rates of change in bone mass depends on the magnitude of change, the precision of the method, and the number of measurements taken.(59) All five bone densitometry techniques are suitable for assessing rates of bone loss, provided that the interval between assessments is sufficiently long. If a technique has too high a precision error, its clinical utility is limited by the length of time it takes to detect significant bone loss in patients.(60,61)
| Technique precision error (CV percent) | Estimated bone loss (percent) | Difference in measurements (percent) b | Approximate followup measurement (years) c |
|---|---|---|---|
| 1 | 1 | 2.77 | 2.77 |
| 1 | 3 | 2.77 | 0.92 |
| 2 | 1 | 5.54 | 5.54 |
| 2 | 3 | 5.54 | 1.85 |
| 3 | 1 | 8.32 | 8.32 |
| 3 | 3 | 8.32 | 2.77 |
| 4 | 1 | 11.08 | 11.08 |
| 4 | 3 | 11.08 | 3.70 |
| 5 | 1 | 13.30 | 13.30 |
| 5 | 3 | 13.30 | 4.43 |
| 6 | 1 | 16.63 | 16.63 |
| 6 | 3 | 16.63 | 5.54 |
a This table assumes that accuracy is invariable.
b Two scans (measurements) would have to differ by more than this amount to be confident that a real change had occurred with 95 percent confidence that the detected losses are real.
c Time frame for a reliable bone mass measurement followup.
Past experience with bone density measurements indicates that strict quality control, including calibration and standardization procedures, is required to maintain both precision and accuracy for reliable measurements.(56,57,62) Because differences in operator technique during scan analysis can result in large measurement variations, it is also important that a consensus be reached over procedural methods for positioning. Those who publish normal ranges should accurately describe the technique used.(63)
Even with instruments calibrated according to manufacturers' instructions, bone density values obtained by DPA and DXA imaging of spine phantoms have differed by as much as 16 percent.(49) Discrepancies in bone density measurements by different instruments appear to reflect differences in instrument design, computer algorithms, and calibration procedures.(58,64,65) For example, when comparing three DXA instruments, Arai et al(65) found each of them to be accurate when using their own standard. However, when the same lumbar spine phantom (standard) was used in each of the three systems, the bone mass density values measured by one system were 5-8 percent higher than those of the other two systems.
The calibration and standardization of bone densitometers is a complex undertaking that requires more attention because there is little agreement among manufacturers. (48,66) Although all manufacturers provide protocols for reference standards and phantom measurements, there is no industrywide agreement for one specific standard that would provide machine-to-machine comparability. Users and/gor manufacturers of bone densitometers need to adopt a uniform procedure for calibrating and standardizing instruments and carry out quality control measurements using an accepted "gold standard."(49)
A report on bone mineral assessment by the National Health Technology Advisory Panel to the Australian Institute of Health recommends an accreditation policy to establish and identify quality assurance programs.(59) A review of the impact of such programs on the precision of densitometric techniques used in multicenter clinical studies shows that a 1-2 percent improvement in precision is achievable.(67)
Single-photon absorptiometry is a noninvasive radiologic technique used to assess bone mass in the appendicular skeleton, usually the radius or calcaneus. It evaluates the amount of bone mass by measuring the transmission of gamma rays through bone. It is a simple, widely available, and relatively inexpensive measurement technique that involves minimum radiation exposure, with negligible bone marrow or gonadal radiation.(59,68)
The earliest SPA measurements were performed at the midradius, which is almost all cortical bone (approximately 95 percent). However, interest in trabecular bone mass, coupled with technical advances in SPA such as rectilinear scanning and area density measurements, provided the capability of assessing sites with greater proportions of trabecular bone: the calcaneus and distal radius.
| Equipment cost | Device: $20,000-$30,000/gsource: $750/4-6 mo |
| Scan charges | $50-$150 |
| Scan site | Distal radius, midradius, calcaneus |
| Scan time | 15 min |
| Measurement units | Mass per unit length of bone (g/gcm) |
| Precision error (CV%)b | Calcaneus and radius (research setting):±1-3%; radius (clinical setting): ±2-5% |
| Accuracy errorc | 3-8% (clinical setting) |
| Radiation source | Current units: 125I/gnewer units: small x-ray tube |
| Radiation exposure | 2-5 mrem |
a Adapted from Health and Public Policy Committee,(68) Lang,(72) Wahner,(64) Johnston,( 73) Mazess.(70)
b Demonstrated by repeat measurements on aluminum standards and patient controls.
c Demonstrated on phantoms and bone specimens.
The accuracy error of SPA measurements is determined from a calibration of the technique with the ash weight of a dried defatted human radius.(74) Calibration can also be made with standards calibrated against ashed bone sections. Single-photon absorptiometry can provide highly accurate measurements (1- 3 percent error) on phantoms and bone specimens. However, accuracy errors of 3-8 percent are more common in the clinical setting.(70)
The precision error of SPA measurements is determined by repeated measurements on phantoms and patient controls. Routine clinical scans of the radius and calcaneus have a precision error of 2-5 percent, largely as a result of repositioning errors and the low and constantly decreasing intensity of the photon source.( 56) Changes in SPA measurements may also occur as a result of treatment that alters body fat; a measured increase in bone density may reflect a decrease in the subcutaneous forearm fat and not a "real" gain in bone.(55) Higher precision measurements are achieved in settings with adherence to quality control and with the application of rectilinear scanning.
The short half-life (60 days) of the 125I source necessitates source changes several times a year and thus has the potential to adversely affect instrument reliability.(56) A similar device that uses an x-ray source (SXA) rather than an isotopic source is also available. The x-ray source, with its significantly greater photon output, improves the precision and the speed of the scans. Because of this and the increasing cost of 125I sources, SPA is being replaced by SXA.
More detailed descriptions of SPA may be found in the 1986 OHTA SPA assessment report by Erlichman(2) and in recent reviews by Lang et al,( 72) Tothill,(75) and Wahner.(64)
Dual-photon absorptiometry is usually used to measure bone density in the spine and hip but can also be used to quantify total body bone mass.(72) It offers no advantage over SPA for measuring appendicular bones. Unlike SPA, its energy source emits photons of two distinct energies, allowing imaging of thicker body parts.
| Equipment cost | Device: $30,000-$65,000/gsource: $5,000-$7,000/12-18 mo |
| Scan charges | $150-$300 |
| Scan site | Spine, hip, total body |
| Scan time | 20-45 min (spine or hip); 40-60 min (total body) |
| Measurement units | Mass per unit area (g/gcm2) |
| Precision errorb | Spine: ±2-5%; femoral neck: ±2-4% |
| Accuracy errorc | Spine: ±3-10%; femoral neck: ±3-8% |
| Radiation source | 153Gd |
| Radiation exposure | 5-10 mrem |
a Adapted from National Health Technology Advisory Panel, (50) Wahner, (64) Mazess, (70) Lang, (72) and Johnston.(73)
b Demonstrated by repeated measurements on controls, phantoms, and ashed bones.
c Demonstrated on phantoms and bone specimens.
Lumbar spine measurements by DPA are usually performed at the L2-L4 region, imaged as a planar area. Dual-photon absorptiometry measurements quantify bone mass of the entire vertebral body, including spinous processes and posterior elements, as well as calcification within surrounding tissues, for example, the aorta. (76) Abnormalities in the spine, including osteophytes, endplate hypertrophy, disc degeneration, calcified aorta, and fractures, are common in patients over 65 years of age.(76,77) These coexisting conditions may result in inaccurate and poorly reproducible vertebral measurements and have led to a preference for femoral measurements.( 76,78)
Dual-photon absorptiometry measurements of the proximal femur are usually performed at the femoral neck but may also include the trochanteric region as well as a highly trabecular area called Ward's triangle.(76) There are few measurement errors attributable to artifacts in the hip region; however, the results can be affected significantly by patient positioning. Dual-photon absorptiometry computer software that does not compensate adequately for the wide variation in the anatomy of the femoral neck and adjacent soft tissues also contributes to inaccurate measurements.
The accuracy error of dual-photon absorptiometry measurements is established by scanning an appropriate spine or femur segment in a cadaver with subsequent determination of the weight of the scanned piece of bone after defatting and ashing. Scanning excised pieces of real bone in water with subsequent defatting and ashing is a less exact procedure for determining measurement accuracy. Another acceptable approach to monitor measurement accuracy uses bone ash or hydroxyapatite in anthropomorphic phantoms with known bone mass and with a known projected bone area.(37) Dual-photon absorptiometry can achieve an accuracy error of about 3 percent on phantoms and bone specimens; however, accuracy errors of 3-10 percent are more common clinically.(54,79)
The precision error of DPA measurements is determined by repeated measurements on phantoms and control patients. Routine clinical scans of the spine and femur have precision errors of 2-5 percent. Imprecision largely results from technical factors (as with all radiologic methods), establishment of suitable soft-tissue baselines, and the low and constantly decreasing intensity of the photon source.(80,81) Higher precision measurements (2-3 percent CV) are achieved with careful attention to baselines, bone edges, and quality assurance.
More detailed descriptions of DPA are found in the 1986 OHTA DPA assessment report by Erlichman(3) and in recent reviews by Wahner,( 64) Mazess,(70) Lang et al,(72) and Tothill.(75)
| Equipment cost | $60,000-$100,000 |
| Scan charges | $150-$300 |
| scan site | Spine and femur |
| Scan time | CELL>5-10 min |
| Measurement units | Mass per unit area (g/gcm2) |
| Precision error (CV)b | ±0.5-3% |
| Accuracy errorc | 3-9% |
| Radiation source | X-ray tube |
| Radiation exposure | < 5 mrem |
b Demonstrated by repeated measurements on phantoms, ashed bones, and control patients.
c Demonstrated on phantoms and ashed bone.
Dual-energy x-ray absorptiometry systems have greater photon output (higher photon flux) than radionuclide sources, improved detector configuration, and automated analysis procedures. These developments provide higher spatial resolution, improved precision, shorter scanning times (10 minutes for the lumbar spine and proximal femur and 20 minutes for a total body scan), and reduced radiation dose compared with QCT and DPA. (63,84- 86) High-resolution imaging facilitates reproducible bone edge detection, improves visualization of the area measured, maximizes the number of bones measurable, and allows artifacts to be recognized and eliminated from the analysis.(66,82) These improvements become particularly important for analyzing spinal bone density in osteoporotics.72,87) Increased scan speed reduces errors resulting from patient motion and is also more convenient for patients with back pain.
Like DPA, DXA measurements reflect the total integrated (cortical and trabecular) mineral and any extraosseous mineral in the path of the beam. These techniques determine the bone density from a conventional anterior-posterior image (AP projection). The sites most commonly measured are the lumbar spine, generally L2- L4 or L1-L4. Other sites include the hip (the femoral neck, the intertrochanteric region, and Ward's triangle) and forearm.(72,88)
Dual-energy x-ray absorptiometry scans of the lumbar spine in the lateral projection were recently developed.(66,89) This may minimize the effect of the posterior arch (cortical bone) on measurements as well as aortic calcification and posterior osteoarthritis, conditions common in the elderly that can falsely indicate elevated bone density.(76) This modification makes possible a more sensitive, direct approximation of trabecular bone.(66) The clinical value of this scanning technique is currently being evaluated.(48,90)
The accuracy error of DXA measurements is determined from a calibration of the technique with ashed bone (excised human vertebrae) or equivalent hydroxyapatite (phantom) in the projected area.(37,84,91) Results are expressed relative to the projected area of bone to give an area-specific bone density in g/gcm2.
Although reviews of recent studies report DXA accuracy error from 3-6 percent, other data indicate that DXA accuracy error on ashed bone specimens is about 9 percent.(91) According to Kelly et al,(82) systematic errors in DXA measurements at the spine can alter the apparent normal range for bone density in the population. They found that the mean value of populations measured by various DXA instruments may differ by as much as 20 percent.
The precision error of DXA measurements is determined by repeated measurements on phantoms and control patients. Routine clinical scans of the spine and femur have a precision error of .5-3 percent.(76, 85,87) These high-precision measurements are largely the result of increased x-ray flux and improved spatial resolution. Long-term precision depends on technician skill for selecting the correct sequence of vertebrae, correct angulation of the leg (femur measurement), and correct bone area. It is essential that the same area be compared each time or additional errors can result.
More detailed descriptions of DXA are available in reports by Wahner et al,( 37) Wahner,(48) Kelly et al,(82) and Kellie.(84)
Quantitative CT is the application of conventional CT scanners and commercially available software packages to the quantitative measurement of bone density in the spine. Dedicated scanners are necessary to obtain satisfactory measurements of bone density. An advantage of QCT, which measures bone in three-dimensional sections (volume), is its ability to separate cortical from trabecular bone with great anatomic detail. Quantitative CT precision has improved (2-6 percent), and both scanning time (15-20 minutes) and radiation dose (about equal to a chest x-ray, but 200-500 times the DXA dose) have been reduced as a result of new calibration standards, software, and improved scanner design.(83)
| Equipment costb | $5,000-$15,000 (software and phantoms) |
| Scan charges | $150-$400 |
| Scan site | Spine |
| Scan time | 10-30 min |
| Measurement units | Mass/gunit volume (mg/cm3) |
| Precision error (CV)c | Dedicated experts: ±1-3%/g common use: ±2-6% |
| Accuracy errord | ±5-15% |
| Radiation source | X-ray |
| Radiation exposure | Older scanners: 500-1,000 mrem; newer scanners: 100-300 mrem |
b To adapt an existing CT scanner.
c Demonstrated by repeated measurements on cadavers, phantoms, and control patients.
d Demonstrated on standard solutions and vertebral specimens.
Different localization and scanning procedures are used to obtain axial sections at the center of the vertebral body. In principle, total, cortical, or trabecular bone density can be measured centrally or peripherally. In practice, however, the method has been developed to measure the central trabecular portion (the spongiosa) of two to four adjacent lumbar vertebral bodies.(66,72) These representative volumes of trabecular bone are quantified and averaged, and the results are expressed as mineral equivalents in g/g cm3. This area covers approximately 35 percent of vertebral trabecular bone and represents a volume of about 3 cm3 of purely trabecular bone in each of the lumbar vertebrae.
With QCT, the change in vertebral (trabecular) bone during the immediate postmenopausal period was reported as 5-7 percent per year, which is about 2-3 times that observed by integral DPA measurements and 3-5 times that seen in the cortical radial sites. Recommendations for monitoring postmenopausal bone loss at the spine with QCT are based on these observations of higher estimates of bone loss. However, overestimation of menopausal spinal bone loss due to variable marrow fat, low precision, and higher radiation exposure have limited this use.(73,75,95)
Although QCT can assess the purely trabecular portion of the spine and may hypothetically provide a more accurate assessment of fracture risk, there is no evidence that this potential benefit outweighs the problems and limitations of the method. Moreover, the contribution of the cortical compartment to crush fracture is being evaluated, and the concept that only the trabecular bone content is important is being challenged.(96,97) No prospective clinical trials with QCT have been conducted to determine its utility for assessing the risk of spinal fractures.
Quantitative CT can be performed with most existing CT units. However, to adapt an existing CT scanner for quantitative measurements of bone density requires software and mineral reference standards (phantoms) with sophisticated calibration and positioning techniques.(78,98) Careful calibration and quality assurance programs are mandatory, especially when the instrument is used both for routine imaging as well as the quantitation of the bone density. (5) However, dedicated scanners are recommended for measurements of bone density. Because CT x-ray output is variable, calibration standards (solutions of varying concentrations of potassium phosphate; K2HPO4) should be measured simultaneously with the patient.(77)
Quantitative CT accuracy is also determined with measurements of vertebral specimens correlated with ash weight. The accuracy error of QCT is 1-2 percent for potassium phosphate solutions and 5-15 percent for human vertebral specimens. A variety of technical factors (scattered radiation, beam hardening, position in the beam, slice orientation) contribute to errors in QCT measurements. Quantitative CT exhibits an accuracy error of 6-9 percent in healthy younger individuals, but becomes increasingly inaccurate (10-15 percent) with aging and with bone disease because of the variability of marrow fat within the vertebral bodies. (70)
The precision error of QCT measurements is determined by repeated measurements on cadavers, phantoms, and patient controls. Some concern remains about the reproducibility of QCT spinal bone density measurements. Despite procedures that have enhanced the technique, some bone studies still show QCT reproducibility to be relatively poor (>3 percent precision error).(99) This is usually because of machine drift and repositioning errors.
The precision error of vertebral QCT in research settings providing exact detail to patient positioning and instrument performance is about 1-3 percent and is 2-6 percent in other settings such as hospitals and clinics.( 48,72,94) However, the precision of multipurpose hospital scanners is generally not better than 5-10 percent.( 83) In research settings with dedicated CT scanners, quality control for bone density measurements is achieved. However, quality assurance in other clinical settings (e.g., hospitals and offices) is frequently lacking.(89,98)
Although most bone mass measurements by QCT use the single-energy scanner, a dual-energy scanner is available that is usually used for research purposes.(100) Although the dual-energy scanner can correct for changes in marrow fat and increase the accuracy of the measurement, it does so at a cost in complexity, precision, and higher radiation exposures.(89,92)
Quantitative CT techniques have been applied on a limited basis to the measurement of trabecular bone at the femoral neck.(72,92) Because of the complexity of the anatomy, reproducible bone density measurements will probably be difficult. Moreover, both cortical and trabecular bone loss may predispose to femoral neck fractures. Therefore, identification of individuals at risk may thus be improved by considering both types of bone deficit. An integrated measurement of cortical and trabecular bone may more reliably predict femoral strength and fracture risk than an independent assessment of either component.( 101)
Ruegsegger et al(102) suggest that cortical bone density and trabecular bone density may complement each other in the diagnosis of bone diseases. They believe that trabecular bone density (radius) is highly sensitive to perimenopausal bone loss or iatrogenic effects of drugs such as corticosteroids. However, the disadvantage of trabecular bone density is the broad range of normal values and its considerable age dependency. In comparison, cortical bone density is limited to a very narrow range and does not show a pronounced age dependency in healthy subjects. An individual deviation of 5 percent from the mean is sufficient to be significant.
A specially constructed CT instrument for the separate determination of trabecular and compact bone density in the appendicular skeleton is available.(77,91,101,102) Although much attention has been given to QCT measurements of lumbar vertebral bodies, a number of factors, including poor precision and accuracy and higher levels of radiation exposure, make this location less than ideal for bone density measurements. In comparison with QCT of the spine, measurements at the radius take less time, have improved precision and accuracy, and expose the patient to less radiation. (103)
| Equipment cost | Existing x-ray equipment/gservice: $60 |
| Scan charges | $90-$160 |
| Scan site | Hand |
| Scan time | 5-10 min |
| Measurement units | Mass per unit area (g/gcm2) |
| Precision error (CV)b | 2-4% |
| Accuracy errorc | 5-10% |
| Radiation source | X-ray |
| Radiation exposure | 10-100 mrem |
b Demonstrated by repeat measurements on phantom hands and patient controls.
c Demonstrated on ashed bones and aluminum standards. Radiographic absorptiometry was developed in an attempt to circumvent some of the inherent limitations and errors in the use of an x-ray image to measure bone mass.
Radiographic absorptiometry involves radiation exposure up to 100 mrem, with negligible bone marrow or gonadal radiation.(68) On average, RA delivers about 40 mrem to the patient.(53) Radiographic absorptiometry bone mass analyses are based on phalangeal scans of pairs of lightly exposed radiographs taken at different kilovoltage and exposure settings. An aluminum alloy reference wedge, which allows the computer to compensate automatically for any between-film differences in exposure, kilovoltage, filtration, film type, or film processing conditions, is included in the image field. A computer-controlled optical densitometer scans the x-ray image with a light beam. The density of the film image over the bone is compared with that over the soft tissue, and the resultant absorption is related to normal values of phalangeal bone obtained from age- and sex-matched controls.(4,107)
Although RA can be used to study the metacarpal, radius, and other appendicular bones, it is most often used to measure the phalanges, a mixture of compact and trabecular bone with minimal overlying soft tissue.(75) X-rays of small bones with minimal soft-tissue cover only require multiplication of the optical densitometry measurement by a constant and do not demand immersion of the hand in a layer of tissue-equivalent material (water bath), according to some investigators.(104) Other researchers do not find variations in soft-tissue thickness corrected by a constant multiplier. (53,109,110) This issue remains controversial.
Radiographic absorptiometry was developed in an attempt to circumvent some of the inherent limitations and errors in the use of an x-ray image to measure bone mass. The accuracy error of RA measurements is determined from a calibration of the technique with ashed bones and an aluminum reference wedge. The wedge serves as a standard in which its measured optical density is compared with the optical densities for bone. Certain technical problems inherent in this technique, however, reduce the accuracy of the measurement.(108) The accuracy error for excised bones is 3-5 percent and 5-10 percent for bones with a slight tissue cover.(53,111) As with the other bone densitometry techniques, correcting the measurement for variation in the amount of soft tissue and fat overlying the bone is a problem. The precision error of RA measurements, determined by repeated measurements on phantom hands and patient controls, is between 2 percent and 4 percent. (112-114)
Radiographic absorptiometry has not been used in any prospective studies of women with osteoporosis or renal osteodystrophy, and very little new information is available in the published literature regarding the clinical utility of RA subsequent to the OHTA assessment report in 1987.(4)
There are those who question whether low bone mass alone is a sufficient explanation for fractures.(10,115,116) Of significance is the discrepancy between bone density and fracture observed in studies in which fluoride caused dramatic increases in the bone density of osteoporotic women without an expected decrease in fractures.(117,118)
Although bone mass correlates well with bone strength, other factors that influence the strength of bone cannot be assessed by bone mass measurements.(119,120) Ultrasound techniques, recently introduced to assess bone mass, can provide information that may be relevant to other aspects of bone quality, such as its geometry, architecture, and biomechanical properties. (121-125 These techniques use either ultrasonic velocity or ultrasonic absorption in bone as an index of bone quality. Preliminary studies(121,123,124) of ultrasound bone assessment are promising and have encouraged interest in the use of ultrasound as a technique to identify women with osteopenia. However, a recent study shows that DXA measurements of the lumbar spine and femoral neck are significantly better than any of the ultrasound findings for determining osteopenia in the hip and spine.(126)
Ultrasound has clinical potential because of its relative low cost and ease of use without the need for ionizing radiation. However, data are too limited to verify the usefulness of ultrasound to assess fracture risk, evaluate bone quality and structure, and identify what characteristics referable to bone quality are being measured.(122)
Biochemical markers of bone turnover (bone formation and resorption) may help to determine the rate of bone loss.(24) There is considerable interest in the use of serum and urinary markers of bone metabolism to predict the skeletal status of women at risk of developing osteoporosis. These biochemical markers can reflect the enzymatic activity of osteoblasts or osteoclasts (e.g., alkaline or acid phosphatase activity); can be proteins synthesized by the bone-forming cells (e.g., osteocalcin or procollagen-1 extension peptides); or can be bone matrix components released into the circulation during resorption (e.g., hydroxyproline or the pyridinoline crosslinks).(35)
According to Arnaud,(127) the utility of some bone markers may be limited by factors that make them both insensitive and nonspecific. Hansen et al,(128) however, predicted future bone mass using SPA measurements of peak bone mass and an estimated rate of bone mass loss based on biochemical measurements. They found that the estimated bone loss was almost identical with the actual loss of bone mass measured 12 years later.
Recently, immunoassays have been developed to quantify urinary excretion of pyridinoline and deoxypyridinoline crosslinks. These were shown to measure the degree of collagen degradation and bone resorption. The potential clinical utility of crosslinks measurement will be in the detection of increased bone turnover in the perimenopause, so that osteoporosis prophylaxis can be instituted before appreciable menopausal bone loss occurs.(127) Alone or in combination with bone absorptiometry, bone markers may aid in the identification of patients at high risk for fracture.(129-131)
End-stage renal disease (ESRD) affected more than 242,000 Americans in 1992, and each year 10,000 to 20,000 new cases are diagnosed.(132,133) Blacks are more often affected than whites, and ESRD is more common in men than in women. Although the survival rate of ESRD patients has improved, metabolic bone diseases that fall under the generic term "renal osteodystrophy" represent abnormal development of bone and major long-term complications.
The association of renal failure with skeletal disease, hyperplasia of the parathyroid glands, and abnormal vitamin D metabolism has been known for many years, and metabolic bone disease leading to mineral loss and clinical symptoms is a well-recognized problem in patients with ESRD.(134,135) Although abnormalities in mineral and bone metabolism occur during the early stages of renal failure, renal osteodystrophy can progress insidiously for many years before symptoms such as bone pain, muscle weakness, fractures, and extraskeletal calcifications occur.(136,137) Treatment of renal osteodystrophy is less effective once symptoms have appeared.
Because ESRD has been associated with bone loss and fractures, bone densitometry techniques have been used to measure BMD in patients with this condition. All the techniques described in Part I (technical report) single- and dual-photon absorptiometry (SPA and DPA), quantitative computed tomography (QCT), radiographic absorptiometry (RA), and dual-energy x-ray absorptiometry (DXA) have been used to measure BMD in patients with ESRD. The purpose of this assessment is to determine whether measuring bone loss associated with uremic disease improves the management of patients with ESRD.
The group of metabolic or uremic bone disorders that characterize renal osteodystrophy includes osteitis fibrosa, osteomalacia, osteosclerosis, osteoporosis, aplastic bone disease, and growth retardation in children; these disorders can occur together in varying degrees.(138,139) The most common forms of renal osteodystrophy are osteitis fibrosa and osteomalacia. Of lesser quantitative importance are osteosclerosis and osteoporosis.
The etiology of these diseases is not completely understood, but impaired vitamin D metabolism, increased concentrations of parathyroid hormone (PTH), diminished gastrointestinal absorption of calcium, phosphate retention, and aluminum toxicity may all play a role.(135,138)
Renal failure results in a variety of biochemical and hormonal derangements. The kidney plays a major role in mineral homeostasis through its capacity to maintain balances for calcium, phosphorus, and magnesium; to synthesize 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] (calcitriol) and 24,25-dihydroxyvitamin D3 [24,25(OH)2D3]; and to degrade and remove PTH from the circulation. Each of these metabolic functions is disturbed to some degree in patients with renal insufficiency.( 140) In addition, certain elements such as lead and aluminum are excreted mainly in the urine, and their retention in the body in patients with renal failure is often associated with clinical disorders of bone and mineral metabolism.
Renal osteodystrophy can produce a variety of histologic changes in bone that are ultimately determined by the biochemical and hormonal abnormalities associated with renal failure. However, there is considerable confusion regarding characterization, nomenclature, and prevalence of the different forms of uremic osteodystrophy.(136) Classification involves a subjective evaluation of data and has varied from two to more than five forms of renal osteodystrophy. Most patients do not present with one clear-cut pathologic condition.
Malluche and Monier-Faugere(141) recommend subdividing renal bone disease into three major histologic groups: predominant hyperparathyroid bone disease, low-turnover uremic osteodystrophy (osteomalacic and adynamic renal bone disease), and mixed uremic osteodystrophy consisting of mild to moderate hyperparathyroid bone disease and defective mineralization. These groups do not represent fully separate entities and transformation from one form to another can occur. Nevertheless, it is useful to distinguish them, because therapy can be tailored according to the predominant histologic findings.
Predominant hyperparathyroid bone disease is characterized by long-standing excessive PTH secretion and a marked increase in bone turnover (high-turnover bone disease). According to Malluche and Monier-Faugere,(141) these characteristics create a particularly fragile bone prone to fractures. This condition is often referred to as osteitis fibrosa. The disease is seen in approximately 5 percent to 30 percent of dialyzed patients.
Low-turnover uremic osteodystrophy is characterized by a striking reduction in the number of bone-forming (osteoblasts) and bone-resorbing (osteoclasts) cells, which result in a significant decrease in bone formation and mineralization.(141) When the reduced mineralization is coupled with a concomitant and parallel decrease in bone formation, the end result is "adynamic uremic bone disease." When diminished mineralization precedes or is more pronounced than the inhibition of collagen deposition, an accumulation of unmineralized matrix is seen as a hallmark of "low-turnover osteomalacia." Osteomalacic bone is prone to deformity, and both types are subject to fractures. This disease is seen in approximately 5 percent to 25 percent of dialyzed patients. Aluminum accumulation in bones has been implicated as a major cause of osteomalacia and aplastic bone disease.(142)
Mixed uremic osteodystrophy lacks a dominant pathogenic basis and is caused primarily by hyperparathyroidism and defective mineralization with or without decreased bone formation. These features may coexist in varying degrees in different patients. This form of renal osteodystrophy is seen in the majority of patients (45 percent to 80 percent) on dialysis.(142)
Aluminum-related bone changes can be seen to varying degrees in all three groups. Approximately 50 percent of unselected patients with chronic renal failure exhibit aluminum deposition in bone, whereas 90 percent of the patients with low-turnover bone disease present aluminum deposits.(142)
| Hypocalcemia | Hypercalcemia |
| Hypophosphatemia | Hyperphosphatemia |
| Raised alkaline phosphatase | Raised serum iPTH |
| Reduced plasma calcitriol | Raised bone aluminum |
| Raised GLA proteinb | Hypermagnesemia |
a Adapted from reference no. 143.
b Osteocalcin.
Abbreviations: GLA = gamma linolenic acid; iPTH = immunoreactive parathyroid hormone.
Phosphate retention plays an important role in producing secondary hyperparathyroidism; (139) however, phosphorus concentrations are generally not indicative of the type or severity of uremic bone disease.(141,144)
The initial tendency to hypocalcemia is attributed to phosphate retention and a reduced concentration of 1,25(OH)2 D.(145) Hypocalcemia leads to PTH secretion, causing parathyroid hyperplasia. Also, hypercalcemia may appear spontaneously in patients undergoing dialysis, in conjunction with an overt secondary hyperparathyroidism and with dialysis osteomalcia.( 135) Like serum phosphorus levels, calcium levels are poor predictors of histologic features in patients with ESRD. Secondary hyperparathyroidism plays a major role in renal osteodystrophy, and raised concentrations of immunoreactive PTH (iPTH) and hyperplasia of the parathyroid glands are among the most consistent abnormalities in patients with chronic renal disease.(137)
Laboratory tests useful in the management of dialysis patients were reviewed in an Office of Health Technology Assessment (OHTA) Report in 1986(146) and revised in 1994.(147)
A common radiographic feature of advanced renal failure is a decrease in BMD, which may arise from secondary hyperparathyroidism, osteomalacia, osteoporosis, or a combination thereof.(135,148) Other radiographic findings seen with some frequency in renal failure are osteosclerosis and fractures. Fractures are more frequent in patients with aluminum osteodystrophy. The presence of upper rib fractures, nonhealing fractures, or three or more nontraumatic fractures is highly suggestive of aluminum-associated bone disease.(141)
Skeletal radiographs should be interpreted in light of the clinical history and biochemical features in order to arrive at a preliminary diagnosis of the specific bone disease. However, normal x-rays do not necessarily exclude the presence of bone disease.(149) Moreover, the interpretation of these radiographic findings has limited clinical value, because radiographic results are not reliable in specifying the type of bone disease present in renal-failure patients.(141,149,150) Although they are of minimal use in the early diagnosis of renal osteodystrophy, radiographs may be useful to locate atraumatic fractures and radiolucent bone cysts in patients undergoing long-term dialysis.(149)
Nearly all patients beginning dialysis have bone abnormalities that become more pronounced and develop into the histopathologic picture of renal osteodystrophy with disease progression.(136) The histologic examination is considered the most effective means of documenting osteitis fibrosa and osteomalacia, as well as identifying aluminum-induced osteodystrophy.
Malluche and Monier-Faugere(136) note, however, that it is not widely used because of the lack of appropriate training of nephrologists, unavailability of specialized laboratories, or reluctance to recommend invasive procedures for diagnosis. Interest in skeletal histology as a guide to the management of renal bone disease has increased substantially.(135,144,151)
A combination of biochemical, clinical, radiographic, and histologic measurements has been used to diagnosis and treat renal osteodystrophy. Therapies range from control of serum calcium and phosphate levels to vitamin D replacement therapy, parathyroidectomy, or aluminum detoxification, depending on the type of bone disease present.(133,134, 152-154)
It remains unclear which tests, other than bone biopsy, are useful in diagnosing the specific pattern of bone disease to aid treatment decisions,(136,149) inasmuch as characteristics that differentiate one bone disease from another frequently do not correlate with biochemical measurements or radiologic features. In addition, laboratory and radiographic abnormalities often do not correspond with the severity of the clinical manifestations.( 140,149)
Because renal osteodystrophy is associated with bone loss, it has been suggested that bone densitometry be used to guide clinical decisions in patients with this disorder. Based on the assumption that the loss of bone associated with renal osteodystrophy predisposes patients to pathologic fractures, the recognition of osteopenia purportedly leads to effective therapy in advance of symptoms and/gor fractures. However, some investigators(149) argue that the low- and high-turnover bone diseases associated with chronic renal failure do not necessarily correlate with BMD and that bone densitometry adds no new diagnostic information to biochemical tests. At issue is whether the detection of bone loss in patients with ESRD effectively alters their medical management. This assessment evaluates the clinical utility of bone densitometry in the management of patients with ESRD and will address such issues as the type and extent of bone loss associated with ESRD and whether these patients have an increased risk for fracture.
Single-photon absorptiometry (SPA), DPA, RA, QCT, and DXA have all been used to measure bone mass in patients with ESRD. The current clinical use of bone densitometry in these patients is based primarily on early SPA studies of cortical bone. Bone densitometry was used to detect early osteopenia and to evaluate its progression, because routine x-rays were considered too insensitive for detecting bone mass loss.(134) Single-photon absorptiometry of the forearm has been the most widely used procedure; other measurement sites possible with SPA include the phalanges, femoral shaft, and ulna, which consist primarily of cortical bone. Controversy regarding the most appropriate measurement site in patients with renal failure and advances in bone densitometry led to studies of BMD in the axial skeleton using QCT at the lumbar spine and DPA and DXA at the spine and hip.
| Reference | Study design/gfollowup | No. of patients | Site | Results: Excess cumulative bone loss compared with age- and sex-matched controls |
|---|---|---|---|---|
| Catto(155) 1973 | Longitudinal/g 6 mo | 13 | Distal radius | 3.8% (mean) |
| Griffiths(156) 1977 | Longitudinal/g 2-5 yr | 195 | Radius and ulna | >= 4 SD |
| Hruska(157) 1978 | Retrospective | 46 | Mid- and distal radius | >= 2 SD (mean) |
| Lindergard(158) 1981 | Longitudinal/g 15-39 mo | 74 | Midshaft radius | 0.84 ± 1.20 SD |
| Seeman(159) 1982 | Retrospective | 14 | Midradius | 1.15 SD (increased)b |
| Muirhead(154) 1983 | Retrospective | 64 | Femoral shaft | >= 1 SD |
| Rickers(161) 1983 | Longitudinal/g 3 yr | 31 | Distal radius | 9% |
| Buccianti(160) 1984 | Longitudinal/g 1 yr | 12 | Radius and ulna | 10% |
| Colbert(162) 1984 | Longitudinalb case control/g2 yr | 355 | Phalangeal bones | 1.6%-3.3% |
| Gupta(163) 1984 | Longitudinal/g13 mo (avg) | 8 | Distal radius | 7% |
a SPA was used in all studies but that of Colbert.(162)
b BMD at the mid-radius was significantly (P <= 0.05) increased.
Abbreviations: RA = radiographic absorptiometry; SD = standard deviation; SPA = single-photon absorptiometry.
| Reference | Study design | No. of patients | Method/glSite | Results: Excess bone mass compared with age- and sex-matched controls |
|---|---|---|---|---|
| Seeman(159) 1982 | Retrospective | 14 | DPA/lumbar spine | +0.11 SD |
| Torres(164) 1982 | Retrospective | 19 | CT/glumbar spine | Normal in 10 patients, high in 5, and low in 4 |
| Piraino(165) 1988 | Retrospective | 31 | CT/glumbar spine | +1.6 SD - osteitis fibros a -1.2 SD - low-turnover osteodystrophy |
| Eeckhout(166) 1989 | Longitudinal/g 3 yr | 20 | DPA/glumbar spine | 17% loss before dialysis 9% gain followup 3 yr on dialysis |
| Eisenberg(167) 1990 | Retrospective | 25 | DPA/glumbar spine hip | 7% loss 35% loss |
| Ito(168) 1991 | Cross-sectional | 16 | CT/glumbar spine | -4.9 to +8.5 SD |
| Chan(169) 1992 | Retrospective case control | 20 | DXA/glumbar spine hip | +0.8% -16.3% |
| Devita(149) 1992 | Retrospective | 30 | DPA/glumbar spine hip | >= 2 SD below the mean >= 2 SD below the mean |
| Gabay(170) 1993 | Cross-sectional | 106 | DPA/glumbar spine hip | -0.06 ± 0.16a -0.97 ± 0.11 |
a Values are mean ± SEM.
Abbreviations: CT = computed tomography; DPA = dual-photon absorptiometry; DXA = dual-energy x-ray absorptiometry; QCT = quantitative computed tomography; SD = standard deviation.
Overall, the studies showed marked differences in the pattern of bone loss: some patients demonstrated gradual or no bone loss, others were "rapid losers" with higher serum iPTH and alkaline phosphatase concentrations,(161) and still others gained bone while on dialysis.(159,166) Rickers et al(161) identified "rapid losers" as patients with bone loss exceeding 10 percent over 3 years, but there was substantial variation in the study group with almost as many patients having a bone loss below 5 percent (slow losers), or no loss at all. Some reports indicated that differences in bone loss were associated with gender and duration of dialysis, whereas others did not.(156,157,161,165,170)
Colbert et al(162) and others(160,169,170) investigated a possible association between the modes of dialysis and the quantity of bone loss due to renal osteodystrophy. Colbert et al(162) demonstrated normal bone loss in ESRD patients on continuous ambulatory peritoneal dialysis (CAPD), Buccianti et al(160) and Gabay et al(170) did not find any significant difference in bone loss between patients on CAPD and on regular hemodialysis, and Chan et al(169) reported equivocal results.
Some studies investigated the relationship between the underlying renal disease process with the quantity of bone loss at the time of dialysis. For instance, Lindergard( 158) found that patients with polycystic kidney disease had lower BMD than patients with glomerulonephritis and that patients with interstitial nephritis tended to have a lower BMD than patients with glomerulonephritis. This could be explained by the longer duration of predialysis uremia in patients with polycystic kidney disease and interstitial nephritis, because these diseases progress more slowly than chronic glomerulonephritis. Similarly, Rickers et al(161) reported that mean BMD loss over 3 years of dialysis was pronounced and significant (P < .02) in patients with chronic pyelonephritis (9.8 percent) and polycystic kidney disease (14.2 percent), but much smaller, and not significant, in patients with chronic glomerulonephritis (4.8 percent).
Although SPA studies have documented cortical bone loss in the context of chronic renal failure, either in predialysis or dialysis patients, there is less, and contradictory, information about axial bone loss in such patients. Three studies( 167,168,170) documented that the femoral neck lost BMD to a much greater extent than did lumbar vertebrae in dialysis patients. In contrast, spinal bone density measurements by DPA and QCT were not always decreased. Seeman et al(159) and Torres and Moya(164) showed spinal preservation of bone mass in patients on dialysis. In fact, three studies reported an increase in BMD at the spine in dialysis patients compared with that in age- and sex-matched controls.(165, 166,169)
Piraino et al(165) used CT measurements to demonstrate differences in vertebral bone density based on the type of bone disease present. They observed increased vertebral BMD in ESRD patients with osteitis fibrosa and decreased vertebral BMD in patients with low-turnover osteodystrophy. Ito et al(168) found spinal QCT measurements of patients undergoing dialysis to be potentially unreliable and misleading and proposed using the QCT axial image of each vertebra to classify patterns of density distribution to evaluate bone changes. This difficulty with QCT measurements is in agreement with Mazess,(171) who also believes that QCT measurements can give rise to misdiagnosis and a faulty evaluation of therapy. However, similar discrepancies have been reported with DPA measurements of spinal bone density.(159,166) Difficulty in achieving accurate spinal BMD measurements has been attributed to bone density losses at axial sites that are not uniform, making local site differences significant, as well as repositioning errors and soft-tissue calcifications.
Data on BMD measurements and on the incidence of fractures in patients with chronic renal failure are limited and contradictory. Piraino et al(165) investigated fracture risk and vertebral bone density measurements in 31 patients on dialysis. In that study, low BMD was not predictive of fracture rate, but the type of renal osteodystrophy was. Moreover, the type of renal osteodystrophy was an important determinant of BMD, and the biochemical measurement, N terminal PTH (N-PTH), was a better predictor for diagnosing the type of bone disease present than was BMD (vide infra). Although they conducted a limited retrospective study, based only on 24 fractures in 8 patients during the course of dialysis (<1 year to >15 years), the authors determined that patients with low-turnover osteodystrophy had a fracture rate of .2 fractures/gdialysis year compared with those with osteitis fibrosa, who had .1 fractures/gdialysis year. Aluminum-associated bone disease, found predominantly in the low-turnover osteodystrophy group, is consistent with the high incidence of rib fractures (10/g14) found in this group. No vertebral fractures and only two hip fractures were reported.
This finding contrasts with reports of "many femoral neck fractures" in dialysis patients,(167,172) but is consistent with the "infrequent occurrence of hip fractures" in dialysis patients reported by Chan et al.(169) Moreover, according to Simpson et al,(173) who based their findings on radiologic features, 87 percent of their patients with fractures in a study group of 70 dialysis patients had osteomalacia. Because low or normal total bone was present in only one-third of the patients with fractures, the authors concluded that osteoporosis as determined radiologically was often not the causative factor. Osteomalacia was suspected as the most important etiological factor. More detailed information regarding fracture risk in patients with chronic renal failure, particularly from prospective clinical trials, is unavailable.
According to Devita et al(149) and others,(142,174) metabolic bone disease can be the result of either increased or decreased bone turnover or mixtures of both, and the specific pattern of bone pathology must be identified before appropriate therapy can be prescribed. In addition to bone density determinations, a number of procedures, such as biochemical measurements, N-PTH levels, skeletal scintigraphy, radiographic analysis of hands and clavicles, deferoxamine stimulation testing (bone aluminum accumulation), and bone biopsy, have been used in the diagnosis of uremic bone disease. (149,165,174)
In some studies,(149,165, 174) comparisons were made to help determine which procedures provided the most useful information for the assessment of bone disease in patients with ESRD. In the only prospective study, Devita et al(149) used these diagnostic procedures (vide supra) to evaluate renal osteodystrophy in 30 patients undergoing dialysis for an average of 5.7 years. Bone biopsies and histology (reference method) were reported for 20 patients and were characterized as secondary hyperparathyroidism alone in 11 patients (high-turnover osteodystrophy), secondary hyperpara-thyroidism with vitamin D deficiency osteomalacia in one patient, and mixed disease (mixed renal osteodys-trophy) in five patients. Aluminum-associated bone disease alone was found in two patients (low-turnover osteodystrophy), and moderate osteoporosis alone was found in one patient.
When bone histomorphometry, DPA bone densitometry, and biochemical data were compared, no significant correlations were found (r values not provided). In addition to DPA, calcium, phosphorus, alkaline phosphates, and x-ray results were not significant predictors of bone biopsy results. None of these procedures could distinguish the nature of the renal osteodystrophy. Devita et al(149) also determined that DPA bone density measurements were normal in 9 of 17 patients with bone disease determined by bone biopsy (sensitivity = 47 percent). A normal bone density did not exclude the presence of bone disease (53 percent false negatives).
The authors(149) found no significant association between decreased bone density and abnormal serum levels of calcium, phosphorus, or alkaline phosphatase. The only biochemical measurement that showed a significant association with a low DPA measurement was serum N-PTH (P = .03). Moreover, an increased concentration of serum N-PTH was highly predictive of the presence of changes consistent with hyperpara-thyroid bone disease (sensitivity = 94 percent). In no case was a normal serum N-PTH level associated with an abnormal DPA study, indicating that bone density measurements did not disclose any unsuspected bone disease.
Because serum N-PTH measurement was considered the most useful test in the initial diagnosis of uremic bone disease, and the deferoxamine stimulation test helped select those patients most likely to have aluminum excess, Devita et al(149) concluded that the combination of these procedures with bone biopsy provided the most efficacious and specific means of diagnosing uremic bone disease.
In the only study to investigate BMD measurements and fractures in dialysis patients, BMD was not predictive of fracture prevalence.(165) It was also demonstrated that the different forms of renal osteodystrophy could not be distinguished by BMD measurements.(149, 165,174) Uremic bone disease may be histologically very active without much loss of BMD and bone density may be normal despite the presence of bone disease.(149) Some patients with ESRD present with deformities and fractures, even though their bone density is normal or high.(136)
Piraino et al,(165) however, demonstrated that the type of renal osteodystrophy present was an important determinant of BMD and fracture prevalence. Moreover, Piraino et al(165) and Devita et al( 149) found that the biochemical parameter N-PTH was a better predictor of the type of renal osteodystrophy than was the BMD determination. A low BMD measurement did not add new information to the diagnosis of suspected bone disease already indicated by an increased concentration of serum N-PTH.
In addition to the limited data, there were also contradictory reports, regarding the association of fracture prevalence with the type of renal osteodystrophy, that require further evaluation. In one study,(141) osteitis fibrosa was referred to as a particularly fragile bone prone to fractures. However, two other studies report that bone fractures are uncommon in patients with osteitis fibrosa.(144,173)
The National Institutes of Health (NIH) have informed the CHCT that end-stage renal osteopenia and osteodystrophy are significant medical problems. However, at the present time, it does not appear that a single bone density technique is capable of capturing the dimensions of bone disease provided by bone biopsy. This assessment has identified the current state of affairs.
The Food and Drug Administration has informed CHCT that the agency has not permitted claims where BMD devices are indicated for use in diagnosing conditions or for indicating therapy.
Metabolic bone disease is a major cause of morbidity in patients undergoing dialysis, and it is generally agreed that monitoring the progression of renal osteodystrophy is essential to the optimal management of patients with chronic renal failure. Monitoring has included biochemical profiles, radiologic examinations, bone density determinations, bone biopsies, and, more recently, an assessment of aluminum accumulation in the body. Excluding bone biopsies, these procedures are of limited diagnostic value, and there is little agreement as to the relative importance of each of these parameters.
Bone biopsy with mineralized histologic tests currently represents the only unequivocal means for diagnosing the type and severity of renal osteodystrophy and aluminum accumulation in bone. However, the invasive nature of the examination precludes its repeated use.
Pure osteoporosis is uncommon in renal disease, where generalized demineralization usually occurs in association with other forms of uremic osteodystrophy. Data on the usefulness of BMD measurements to diagnose uremic bone disease or to select treatments in patients with ESRD are limited. Moreover, bone densitometry is unable to provide specific diagnostic information in the assessment of uremic bone disease. Although bone loss has been demonstrated in patients with renal failure, evidence supporting the clinical utility and appropriate role for BMD measurements by any technique at any site in the management of patients with renal failure is lacking.
Bone mass density determinations by SPA and RA in patients with ESRD have been included in the Medicare instructions for ESRD since the early 1980s. However, the results of this analysis suggest that BMD measurements are not able to differentiate uremic bone diseases or predict fracture risk in patients with renal osteodystrophy. As a result, BMD measurements currently do not provide useful information that could support therapeutic decisions in the management of these patients.
AP = Anterior-posterior
BMD = Bone-mass density
CAPD = Continuous ambulatory peritoneal dialysis
CDC = Centers for Disease Control and Prevention
CHCT = Center for Health Care Technology
CT = Computed tomography
DPA = Dual-photon absorptiometry
DXA = Dual-energy x-ray absorptiometry
ESRD = End-stage renal disease
FDA = Food and Drug Administration
g/gcm = grams per centimeter
g/gcm2 = grams per square centimeter
g/gcm3 = grams per cubic centimeter
HCFA = Health Care Financing Administration
iPTH = Immunoreactive parathyroid hormone
NIH = National Institutes of Health
N-PTH = N-terminal parathyroid hormone
OHTA = Office of Health Technology Assessment
PHS = Public Health Service
PTH = Parathyroid hormone
QCT = Quantitative computed tomography
RA = Radiographic absorptiometry
SD = Standard deviation
SPA = Single-photon absorptiometry
SXA = Similar device that uses an x-ray source
WHO = World Health Organization
