Radiographic Absorptiometry as a Diagnostic and Management Tool for Osteoporosis

Osteoporosis

Osteoporosis is a disease characterized by deterioration of the strength and fracture resistance of bone due to decreased bone mineral density. Although osteoporosis affects members of both sexes, more is known about the disease in women than in men. Osteoporosis affects 32% of white, postmenopausal women in Minnesota, 75% of whom have osteoporosis of the hip. Each year, approximately 3,700 Minnesota women fracture their hips. Discussion is confined to white women due to limited research on osteoporosis in men, as well as women in different racial and ethnic groups.

Of all osteoporosis-related fractures, hip fractures pose the greatest public health problem. Nationally, up to 20% of women who fracture a hip die within one year; about half of women living at home at the time of a hip fracture experience decreased social functioning within 2 1/2 years. The costs of caring for a woman with a hip fracture were estimated at $19,335 per patient in 1990 in the United States. These costs included in-hospital services, post-hospital nursing home care, and outpatient services such as home health care.

A woman is considered at risk for osteoporosis if she is Caucasian or Asian, has a small frame, is short and light weight, has lower than normal estrogen levels, has experienced early menopause, does not get enough calcium or vitamin D in her diet, uses caffeine, tobacco or steroids, and does not exercise sufficiently.

Osteoporosis is an asymptomatic disease. Often, a woman and her health care provider only find out she has the disease when a fracture occurs. However, once a woman has an osteoporosis-related fracture her treatment options are limited. Hormone replacement therapy (HRT), which can consist of estrogen, progesterone, or both, and alendronate (brand name Fosamax) are approved by the Food and Drug Administration (FDA) as preventives and treatments for osteoporosis. Both drugs slow bone loss and increase bone mass. Another drug, calcitonin slows bone loss, but does not increase bone mass. Of the three, HRT is the most widely used treatment for osteoporosis. This is because it decreases a woman's risk of fracture in addition to decreasing her risk for heart disease. Many women refuse HRT, however, because it is associated with an increased risk of breast and endometrial cancers. Unlike HRT, alendronate and calcitonin do not appear to increase a woman's risk of breast or endometrial cancer. However, in the case of alendronate there have been reported cases of esophageal ulceration when it is incorrectly taken. Up to 20% of patients who take calcitonin experience side-effects, including nausea. Both calcitonin and alendronate lack the cardio-protective effect of HRT.

Although prevention remains a woman's best defense against osteoporosis, availability of new treatments for the disease has increased patient and practitioner interest in access to bone densitometry services.

Bone Densitometry

A woman's bone mineral density (BMD) is used as an indicator that she has osteoporosis or is at increased risk for the disease. The World Health Organization's criteria, which defines osteoporosis in terms of bone mineral density, is commonly used by practitioners in the diagnosis of the disease. A woman who is more than 2.5 standard deviations below a healthy, 30 year old woman's bone mass is considered to have osteoporosis. A woman is considered osteopenic when her bone density falls between 1 to 2.5 standard deviations below this norm. Usually, a woman reaches peak bone mineral density by age 30. Her BMD declines thereafter as her body's ability to replace old or damaged bone with healthy new bone decreases. This process, called resorption, accelerates after menopause due to decreased estrogen levels.

Two bone mineral density measurement techniques are used in Minnesota; quantitative computed tomography (QCT) and dual energy x-ray absorptiometry (DEXA). DEXA is the most commonly used method. As of October 1997, there were 30 stationary sites using DEXA, most concentrated in Minneapolis, St. Paul and Rochester, and 4 stationary QCT sites. In addition, four companies are currently operating mobile DEXA units in the state. Three are based in Minnesota and the other is based in Sioux Falls, South Dakota. As of October 1997, 45 sites in Greater Minnesota are regularly visited by these mobile units. A map showing the availability of DEXA and QCT by city and county in Minnesota is in available upon request. The map also includes mobile DEXA sites on the border between Minnesota and North and South Dakota.

Charges for bone densitometry vary depending on the body site being measured and the technique being used. Average charges are $150 to $300 for DEXA of the hip or spine, $150 to $400 for QCT of the spine and $60 to $90 for RA of the fingers. Medicare will reimburse providers of bone densitometry for one BMD test per year per patient. Currently, the Medicare reimbursement rate is $131.00 for DEXA of the hip or spine and $51.00 for RA of the hand. Medicare does not currently reimburse for peripheral DEXA (pDEXA) of the forearm or hand. For RA, charges include the x-rays and their analysis by the OsteoGram Analysis Center. Capital costs for DEXA equipment are $60,000 to $100,000. QCT is less costly since specialized software is used to adapt an existing CT scanner. The software costs between $5000 and $15,000. By contrast, no capital outlay is required for RA since existing x-ray equipment is used. In Minnesota, there are approximately 570 x-ray facilities with one or more x-ray machines on site. These facilities may potentially use their equipment to perform RA.

RA's only provider in the United States, the OsteoGram Analysis Center, has received exemptions for their procedure from a number of states. These states have exempted RA based on their determination that RA is not "a routine diagnostic test" or will not be used for "routine imaging."

Background Information

Radiographic absorptiometry (RA) is a method of determining bone mineral density (BMD) from x-rays (plain radiographs) for the purpose of detecting osteoporosis. This technique is typically applied to peripheral measurement sites such as the bones between the wrist and fingers (metacarpals) or the fingers (phalanges) of the hand. The RA technology available in the United States was developed by CompuMed, Inc (Manhattan Beach, CA), and licensed by Merck and Co., Inc. RA assesses bone mineral density by taking two radiographs of the left hand using direct exposure (nonscreen) x-ray film at different radiographic settings. An aluminum alloy wedge is placed in a specific location next to the hand as a reference for computerized analysis of bone image optical density data. The parameters calculated from the reference wedge are applied to the bone image data from the middle phalanges of the index, third, and fourth fingers, and bone mineral mass is computed. Bone volume is estimated, based on the assumption that the bones are cylindrical in shape, and BMD is calculated. Corrections for soft tissue absorption, small errors in subject positioning, and variables associated with the film or exposure technique are incorporated into the analytical algorithms. Interpretation of the results is based on comparison with a set of reference data. Computerized analysis of the radiographs is performed by the OsteoGram Analysis Center. Although no other centers in this country currently offer RA analysis, the technique is available in Europe through the use of Osteoradiometer (marketed by NIM, Verona, Italy). Modified forms of RA are performed in Japan with the Bonalyzer (Teijin Co, Ltd, Tokyo, Japan), referred to as computer-assisted x-ray densitometry (CXD), and in the Netherlands with a technique known as quantitative microdensitometry (QMD).

Several other techniques used to detect osteoporosis also employ the use of plain nonscreen radiographs. Plain radiographs are used for radiogrammetry, or computed radiogrammetry, techniques that use measurement of the external and internal diameters of the metacarpal bone cortex to distinguish normal from osteoporotic bone. Several Japanese studies have reported on the use of a photodensitometric technique using nonscreen radiographs to quantify bone mass.

A variety of other techniques are used to diagnose osteoporosis. Dual energy x-ray absorptiometry (DEXA) requires specialized equipment, can be performed on the axial (e.g. hip or spine) or appendicular (e.g., forearm, fingers, or hand) skeleton, and is considered by many to be the gold standard for detecting osteoporosis. Quantitative computed tomography (QCT), a three-dimensional technique which can measure true BMD volume, requires a CT scanner and specialized software. Nonradiographic techniques under investigation involve the use of ultrasound (US). However these are not approved by the Food and Drug Administration (FDA) for measurement of BMD at the present time.

Clinical Research Studies

Evaluation of RA as a diagnostic tool involves several issues. The first issue is whether RA can accurately detect osteoporotic bone in the region of interest, usually the hand or forearm (distal radius). The degree of accuracy is determined by the precision, sensitivity, and specificity of the technique. The second issue is the predictive value of RA. This question involves whether the presence of osteoporosis in a peripheral skeletal site, such as the fingers, hand or forearm, is a good predictor of osteoporosis in areas with clinical and social relevance for fractures, such as the vertebrae and hip. A limited number of studies assessing the validity of RA for determining BMD are available; most are comparison studies using DEXA as the gold standard. Among these studies, there are often differences in study populations, differences in the definition of osteoporosis, or differences in the algorithms used to differentiate cortical bone from soft tissue or to assess the density of bone. In addition, studies may report different parameters, including cortical thickness, bone mineral content (BMC) or BMD. Most studies involve relatively small numbers of subjects. Moreover, because very few reports involve longitudinal studies it is difficult to assess the ability of RA measurements to predict vertebral or hip fracture.

A summary of the published studies assessing RA for diagnosis of osteoporosis; studies using modified RA techniques, such as QMD and CXD is available. A variety of statistical measures were used in the reviewed studies to compare techniques and to assess the precision and accuracy of RA for diagnosis of osteoporosis and RA's predictive value for future osteoporosis-related fractures. Determination of correlation coefficients (r) was used as a test of agreement between RA results and those of DEXA, QCT, and other techniques. Calculation of kappa (k) scores was also used to determine the diagnostic agreement among the various techniques. A measure of precision, the coefficient of variation (CV), was determined in some studies for intra-assay (same observer measuring same film multiple times), interassay (same observer measuring different films), and interobserver (different observers measuring same film) variability. Receiver operating characteristic (ROC) curve analysis is a graphic means for assessing the ability of a screening test to discriminate between healthy and diseased persons, and the area under the curve (AUC) gives an indication of the sensitivity of the test as specificity varies, such as with different cutoff points. Odds ratios were used to quantify and compare the ability of RA and DEXA to discriminate between healthy and osteoporotic women. In one prospective study, Cox proportional hazards models were constructed to provide age-adjusted estimates of the risk of sustaining a fracture according to BMD or bone loss, as measured by QMD.

In a study of a subgroup of early postmenopausal women enrolled in a multicenter, longitudinal, interventional cohort study sponsored by Merck, Sharp and Dohme, Ravn et al. (1996) compared RA measurements of BMD with single-energy x-ray absorptiometry (SXA) of the distal forearm, and DEXA of the lumbar spine, proximal femur, and distal forearm. Information from this study indicated that the precision error was 1.5% for RA and 1.0%-2.2% for SXA and DEXA. BMD measured by RA correlated with BMD measured by SXA and DEXA in the range 0.46 to 0.72 (p less than 0.001). These data suggest that RA of the phalanges has a precision error similar to that of SXA and DEXA, and can have a high correlation with these techniques.

While the studies discussed have addressed the comparative agreement of RA measurements of BMD with other bone densitometry techniques, little information is available regarding the predictive value of these measurements for clinically and socially relevant problems, such as vertebral and hip fractures. Although RA was not used to measure BMD, Cummings et al. (1993) concluded that BMD measurements had value for prediction of hip fractures. In a multicenter, longitudinal cohort study involving 8134 women, 65 women sustained confirmed hip fractures. Women with BMD in the lowest quartile had an 8.5-fold greater risk of hip fracture than those with BMD in the highest quartile. Although BMD of the femoral neck was the best predictor, measurements of the radius and calcaneus also had predictive value for hip fracture.

In a collaborative study with researchers from CompuMed, Inc and Merck and Co, Inc, Mussolino et al. (1997) assessed the long-term predictive utility of RA measurements of phalangeal BMD for hip fracture risk. Data for the analysis in this study were from a longitudinal cohort study (First National Health and Nutrition Examination Survey) and the Epidemiologic Follow Up Study, in which x-ray examinations of the hand had been made on a subsample of the cohort (n=6913). RA analysis of these radiographs was performed for 3481 subjects. Calculation of the relative risk for hip fracture indicated that RA measurements of BMD and bone mass were significant predictors of future hip fracture risk. The relative risk values ranged from 1.57 to 1.81 depending on the specific survival model used for the analysis.

Conclusions

RA appears to provide measurements of BMD comparable to those of other bone densitometry techniques. Evidence from the literature suggests that RA measurements of BMD have moderate to good correlation with measurements made using other techniques, and can detect osteoporosis of the fingers, hand or forearm with a similar degree of precision. The highest correlations between measurements of BMD among the various techniques are with measurements taken at the same body site. For example, RA of the distal forearm, hand and/or fingers appears to provide similar information as DEXA of the distal forearm. However, the predictive value of BMD measurements for vertebral and hip fractures is less clear. Prospective studies to address this question have not yet been done. It can be inferred from the available studies that RA measurements of BMD have some value in prediction of hip and vertebral fractures. Nevertheless, BMD measurements of the fingers, hand or forearm by RA do not predict future hip or vertebral fractures as well as BMD measurements of the hip or spine, respectively.

General Procedure

The OsteoGram Starter Kit instructs the operator to take two pictures of the left hand. A readipack, which consists of unexposed x-ray film enclosed in a manila packet, is used. A template showing the correct positioning of the hands and the aluminum reference wedge is printed on the outside of the film packet. A 10 kilovolt (kV) separation should be used between the two exposures. The kit recommends use of the enclosed readipacks because they are designed for direct exposure x-ray radiation. However, according to the Procedures Manual, other double emulsion film sheets may be used with the cardboard holder and template included in the kit. The OsteoGram Procedures Manual recommends an x-ray background between 0.9 and 1.5 optical density, although an optimal level is 1.1 optical density. A reference picture is included in the kit so the provider or technologist may compare their OsteoGram's background optical density against it.

Findings

Dr. Payne performed two tests. In the first test, he took the readipack included in the OsteoGram Starter Kit and exposed it to radiation following the Procedures Manual instructions. Using a high frequency x-ray system, Dr. Payne got a background of 1.16 optical density at a setting of 51 kV, 49 milliamp seconds (mAs) at the recommended 40 inch distance between the x-ray tube and film. The film was processed using a Konica processor. Dr. Payne measured the radiation exposure to the readipack film as 95 milliroentgen (mR). As a second test, Dr. Payne took the cardboard cassette holder and inserted a nonscreen, double emulsion, medium speed Konica film commonly used in his lab. It does not have as much emulsion as the Kodak readipack film, so he had to go up to 194 mAs to get an optical density of 1.03, and it took 380 mR of radiation to expose the film. As a point of comparison, to take an x-ray of the hand, a single emulsion film with a detail, or intensifying screen, a setting of 60 kV and 18 mAs would result in a radiation exposure of 15 mR to the patient.

Since the OsteoGram process requires two radiographs be taken, radiation exposures will range between 180-200 mR, using a readipack, up to 800 mR using a double emulsion film in a cardboard cassette holder.

Conclusions

According to Dr. Payne, occupational safety limits are set by the National Council on Radiation Protection (NCRP) and International Council on Radiation Protection (ICRP). Neither the NCRP or the ICRP address radiation safety limits for medical patients or dose exposure to individual organs. An averaging process called whole-body equivalent dose or effective dose equivalent is used as a measure of the health risks from ionizing radiation. Occupational risks are assessed primarily in terms of radiation exposure that can cause fatal cancers. Dr. Payne conducted a search of the medical literature and found nothing on hand cancer.

In Dr. Payne's opinion, exposing the hand to 15 mR, 100 mR or 800 mR does not pose a health threat. He also did not attach significance to scatter radiation to other parts of the body. However, use of the readipack does expose the individual to less radiation than other types of x-ray film (95 mR versus 380 mR). The nationally accepted principle of ALARA ("as low as reasonably achievable") dictates that an individual be exposed to the least amount of radiation necessary to obtain good diagnostic information. However, film that uses an intensifying screen to decrease radiation exposure further cannot be used with the OsteoGram test process. The image produced by such x-ray film cannot be analyzed by the current OsteoSystem software program. If a physician wants to use this process, he/she must use x-ray film without an intensifying screen.

The work group was informed by the Minnesota Department of Health Radiation Control Section that the Kodak readipack used by OsteoGram is no longer being manufactured as of September 1997. However, inventory of the Kodak readipack is still available and OsteoGram is trying to find another vendor to provide them with readipack film.

Whether there is a limit to the number of hand x-rays someone should be exposed to in a year, or a lifetime, is unclear. The NCRP has recommended radiation exposure to the hand be limited to 50,000 mR annually for a radiation worker. Whole body, or effective dose equivalent, radiation exposure for workers is limited to 5,000 mR annually. Even if a woman had RA of the hand five times in one year at the maximum exposure of 800 mR, she still would be within acceptable radiation exposure limits. However, it is unlikely that five hand RA's in one year would be done. Bone density changes slowly. A practitioner would get no relevant information from doing it more often than once a year. In addition Medicare pays for only one exam per year for patients at risk for osteoporosis.

Given the radiation levels described, it was Dr. Payne's opinion, that access to RA by physicians to identify women at risk for osteoporosis or osteopenia does not pose a health threat to women in Minnesota. However, current state regulation of direct exposure x-ray film is based on the ALARA principle. Dr. Payne indicated that to do the RA test procedure correctly, with the minimum amount of radiation, the readipack should be used and another source for the readipacks should be pursued. The other x-ray film he used in his second test is designed to be used with an intensifying screen. As a practitioner, Dr. Payne would prefer exposing the patient to 100 mR rather than 400 mR. But, it is not possible to further reduce the radiation that the patient is exposed to from 100 mR down to 15 mR; it would not be the same diagnostic test. In addition, Dr. Payne stated that although DEXA exposes the patient to significantly less radiation (39-77 mR), 100 mR for any diagnostic test is probably not significant.

Finally, low dose exposure to a central body site versus a higher dose exposure to a peripheral body site is not the primary consideration when choosing a diagnostic test. Dr. Payne indicated that the important question is not how much radiation is delivered, but which test provides the best diagnostic information.

Effectiveness and Predictive Value of RA

It appears from the available evidence that RA does provide diagnostic information comparable to that obtained by using other bone densitometry techniques. RA measurements of the forearm, hand and/or fingers correlate well with measurements of the forearm using DEXA. Peripheral site measurements using RA correlate only moderately with measurements taken using other bone densitometry techniques of non-peripheral sites, such as the hip and spine. As is the case with all bone densitometry techniques, RA's ability to predict future hip or vertebral fractures based on BMD measurements is undetermined. It should also be noted that RA should not be used if an individual has osteoarthritis or other hand deformity, which makes it difficult to lay the hand flat on the readipack.

Safety of RA

During the course of work group meetings, the following perspectives on RA's safety emerged.

• RA's radiation levels are well within safe limits. A woman being tested using RA will be exposed to approximately 200 to 800 mR annually, depending on the type of x-ray film used. This level of radiation exposure is necessary to take an image of good diagnostic quality that can be analyzed by the OsteoGram software system.
• The nationally accepted ALARA principle dictates that individuals be exposed to the minimum amount of radiation necessary to obtain an image of good diagnostic quality. Other states' rules and regulations, including Colorado, Delaware, Nevada, Washington, Tennessee, Arizona, and South Carolina, require that diagnostic tests use the minimum radiation exposure necessary to produce images of good diagnostic quality. Many states require use of intensifying screens and the highest speed film available consistent with diagnostic objectives. (A summary of state rules and regulations is available.) However, radiation exposure to the individual can be minimized in RA if the readipack designed for direct exposure x-rays is used rather than x-ray film designed for use with an intensifying screen. The high end of the dose range noted above for RA is the result of using film which is designed to be used with intensifying screens.
• Although the radiation dose from DEXA is lower, the areas being exposed (the hip or spine) may be at greater risk for disease because soft tissue organs are also being exposed to ionizing radiation. RA only exposes the hand and fingers to ionizing radiation. However, the NCRP and ICRP do not specify what limits should be put on patient exposure to radiation in a medical setting, nor do they address radiation exposure levels to individual organs.

Availability of Bone Densitometry Services and RA's Potential Impact on Individuals in Minnesota

The availability of RA has the potential to increase access to bone densitometry services in Minnesota. Although mobile DEXA units have increased the availability of bone densitometry to rural Minnesotans, coverage is incomplete and contractual arrangements between the mobile service provider and the site provider may vary. A mobile provider may contract to visit once every two weeks or once a month. Contracts are negotiated because of provider interest and because the number of patients justifies the time and expense required to provide mobile bone densitometry services. Further, access to care remains an issue. Travel to the mobile unit's scheduled site visit may be necessary in order for an individual to receive a BMD test. If distances are great enough, this might discourage him or her from getting the test. By contrast, RA can be performed on site, requiring only an x-ray machine and an OsteoGram Starter Kit. It is readily accessible. The trade-off is increased exposure to ionizing radiation. However, if RA is the only test available to make the diagnosis of osteoporosis, the benefits may outweigh the risks.