|Home | About | Journals | Submit | Contact Us | Français|
Systematic reviews and analyses of administrative data were performed to determine the appropriate use of bone mineral density (BMD) assessments using dual energy x-ray absorptiometry (DXA), and the associated trends in wrist and hip fractures in Ontario.
Dual energy x-ray absorptiometry bone densitometers measure bone density based on differential absorption of 2 x-ray beams by bone and soft tissues. It is the gold standard for detecting and diagnosing osteoporosis, a systemic disease characterized by low bone density and altered bone structure, resulting in low bone strength and increased risk of fractures. The test is fast (approximately 10 minutes) and accurate (exceeds 90% at the hip), with low radiation (1/3 to 1/5 of that from a chest x-ray). DXA densitometers are licensed as Class 3 medical devices in Canada. The World Health Organization has established criteria for osteoporosis and osteopenia based on DXA BMD measurements: osteoporosis is defined as a BMD that is >2.5 standard deviations below the mean BMD for normal young adults (i.e. T-score <–2.5), while osteopenia is defined as BMD that is more than 1 standard deviation but less than 2.5 standard deviation below the mean for normal young adults (i.e. T-score< –1 & ≥–2.5). DXA densitometry is presently an insured health service in Ontario.
The Canadian Multicenter Osteoporosis Study (CaMos) found that 16% of Canadian women and 6.6% of Canadian men have osteoporosis based on the WHO criteria, with prevalence increasing with age. Osteopenia was found in 49.6% of Canadian women and 39% of Canadian men. In Ontario, it is estimated that nearly 530,000 Ontarians have some degrees of osteoporosis. Osteoporosis-related fragility fractures occur most often in the wrist, femur and pelvis. These fractures, particularly those in the hip, are associated with increased mortality, and decreased functional capacity and quality of life. A Canadian study showed that at 1 year after a hip fracture, the mortality rate was 20%. Another 20% required institutional care, 40% were unable to walk independently, and there was lower health-related quality of life due to attributes such as pain, decreased mobility and decreased ability to self-care. The cost of osteoporosis and osteoporotic fractures in Canada was estimated to be $1.3 billion in 1993.
With 2 exceptions, almost all guidelines address only women. None of the guidelines recommend blanket population-based BMD testing. Instead, all guidelines recommend BMD testing in people at risk of osteoporosis, predominantly women aged 65 years or older. For women under 65 years of age, BMD testing is recommended only if one major or two minor risk factors for osteoporosis exist. Osteoporosis Canada did not restrict its recommendations to women, and thus their guidelines apply to both sexes. Major risk factors are age greater than or equal to 65 years, a history of previous fractures, family history (especially parental history) of fracture, and medication or disease conditions that affect bone metabolism (such as long-term glucocorticoid therapy). Minor risk factors include low body mass index, low calcium intake, alcohol consumption, and smoking.
The Ontario Health Insurance Program (OHIP) Schedule presently reimburses DXA BMD at the hip and spine. Measurements at both sites are required if feasible. Patients at low risk of accelerated bone loss are limited to one BMD test within any 24-month period, but there are no restrictions on people at high risk. The total fee including the professional and technical components for a test involving 2 or more sites is $106.00 (Cdn).
This review consisted of 2 parts. The first part was an analysis of Ontario administrative data relating to DXA BMD, wrist and hip fractures, and use of antiresorptive drugs in people aged 65 years and older. The Institute for Clinical Evaluative Sciences extracted data from the OHIP claims database, the Canadian Institute for Health Information hospital discharge abstract database, the National Ambulatory Care Reporting System, and the Ontario Drug Benefit database using OHIP and ICD-10 codes. The data was analyzed to examine the trends in DXA BMD use from 1992 to 2005, and to identify areas requiring improvement.
The second part included systematic reviews and analyses of evidence relating to issues identified in the analyses of utilization data. Altogether, 8 reviews and qualitative syntheses were performed, consisting of 28 published systematic reviews and/or meta-analyses, 34 randomized controlled trials, and 63 observational studies.
A systematic review showed that the mean rate of bone loss in people not receiving osteoporosis treatment (including postmenopausal women) is generally less than 1% per year. Higher rates of bone loss were reported for people with disease conditions or on medications that affect bone metabolism. In order to be considered a genuine biological change, the change in BMD between serial measurements must exceed the least significant change (variability) of the testing, ranging from 2.77% to 8% for precisions ranging from 1% to 3% respectively. Progression in BMD was analyzed, using different rates of baseline BMD values, rates of bone loss, precision, and BMD value for initiating treatment. The analyses showed that serial BMD measurements every 24 months (as per OHIP policy for low-risk individuals) is not necessary for people with no major risk factors for osteoporosis, provided that the baseline BMD is normal (T-score ≥ –1), and the rate of bone loss is less than or equal to 1% per year. The analyses showed that for someone with a normal baseline BMD and a rate of bone loss of less than 1% per year, the change in BMD is not likely to exceed least significant change (even for a 1% precision) in less than 3 years after the baseline test, and is not likely to drop to a BMD level that requires initiation of treatment in less than 16 years after the baseline test.
There are presently no specific Canadian guidelines for BMD screening in men. A review of the literature suggests that risk factors for fracture and the rate of vertebral deformity are similar for men and women, but the mortality rate after a hip fracture is higher in men compared with women. Two bisphosphonates had been shown to reduce the risk of vertebral and hip fractures in men. However, BMD testing and osteoporosis treatment were proportionately low in Ontario men in general, and particularly after a fracture, even though men accounted for 25% of the hip and wrist fractures. The Ontario data also showed that the rates of wrist fracture and hip fracture in men rose sharply in the 75- to 80-year age group.
The economic analysis focused on analyzing the economic impact of decreasing future hip fractures by increasing the rate of BMD testing in men and women age greater than or equal to 65 years following a hip or wrist fracture. A decision analysis showed the above strategy, especially when enhanced by improved reporting of BMD tests, to be cost-effective, resulting in a cost-effectiveness ratio ranging from $2,285 (Cdn) per fracture avoided (worst-case scenario) to $1,981 (Cdn) per fracture avoided (best-case scenario). A budget impact analysis estimated that shifting utilization of BMD testing from the low risk population to high risk populations within Ontario would result in a saving of $0.85 million to $1.5 million (Cdn) to the health system. The potential net saving was estimated at $1.2 million to $5 million (Cdn) when the downstream cost-avoidance due to prevention of future hip fractures was factored into the analysis.
There is a lack of standardization for BMD testing in Ontario. Two different standards are presently being used and experts suggest that variability in results from different facilities may lead to unnecessary testing. There is also no requirement for standardized equipment, procedure or reporting format. The current reimbursement policy for BMD testing encourages serial testing in people at low risk of accelerated bone loss. This review showed that biannual testing is not necessary for all cases. The lack of a database to collect clinical data on BMD testing makes it difficult to evaluate the clinical profiles of patients tested and outcomes of the BMD tests. There are ministry initiatives in progress under the Osteoporosis Program to address the development of a mandatory standardized requisition form for BMD tests to facilitate data collection and clinical decision-making. Work is also underway for developing guidelines for BMD testing in men and in perimenopausal women.
At the request of the Ontario Health Technology Advisory, the Medical Advisory Secretariat conducted a review of the utilization of and evidence on bone mineral density (BMD) testing for the identification and diagnosis of osteoporosis.
Bone mineral density testing has been used to detect and diagnose osteoporosis, a systemic skeletal disease characterized by low bone mass and micro-architectural deteriorations of bone tissue, with a consequent decrease in bone strength and increased susceptibility to fracture (1)
Extrapolating the age-specific rates to Canada, Goeree et al. (2) estimated that in 1993, approximately 1.8 million Canadian females had osteoporosis, particularly postmenopausal and elderly women. The Canadian Multicentre osteoporosis Study (CaMos) estimated that 16% of Canadian women and 6.6% of Canadian men have osteoporosis as defined by the World Health Organization. Osteopenia was found in 49.6% of Canadian women and 39% of Canadian men. (3)
Osteoporosis predisposes individuals to fragility (low trauma) fractures, defined as fractures relating to a fall from the standing position. (4) Fracture sites most likely to be associated with osteoporosis are the pelvis, spine, wrist, proximal femur, and proximal humerus. Vertebral fractures in the thoracic or lumbar spine are highly suggestive of osteoporosis, but are frequently undetected because they are asymptomatic. Colle’s fracture appears 10 years before hip fracture, and is a determining factor for hip fracture. (5;6)
Recent data suggest that approximately 3% of Canadians over the age of 25 sustain a fragility fracture each year, with the majority of serious fractures occurring in people over age 50 years. The risk of fragility fractures is particularly high in women. In many Western countries, the remaining life-time risk of a hip fracture in white women at the age of menopause was estimated to lie between 15 and 17%, with the remaining life-time risk for all fractures reaching 30 to 40%. (7) Fractures occurring at the spine and the forearm are associated with significant morbidity, while hip fractures are associated with significant increase in mortality. (7) Papadimitroupoulos et al. (8) reported in 1997 that the incidence of hip fracture and death rates during acute hospitalization in Canada increased exponentially with increasing age, and projected an increase in the number of age-adjusted hip fractures from 23,375 in 1993-1994, to 88,124 in 2041. A similar trend was observed for the province of Ontario. Jaglal et al. (9) reported in 1996 that between 1981 and 1992, the overall hip fracture rate in Ontario (based on hospital discharge data) was 3.3 per 1,000 persons (4.6 per 1,000 women vs 1.7 per 1,000 men). With the aging of the population, Jaglal et al. (9) projected that the number of hip fractures in Ontario would double by 2010, and that hospital bed-days due to hip fracture would increase by 84%, from 214,000 in 1990 to 393,000 in 2010.
Osteoporotic fractures in the elderly population are associated with higher mortality than in the general population (10;11) The mortality rate reaches 20% in the year after a hip fracture, and 20% of the survivors will eventually require long-term care in an institution. (2) A 2001 Ontario study reported that among community dwelling people who had a hip fracture, only 59.4% resided in the community 1 year following a hip fracture, and 5.6% of people who survived their first fracture experienced a subsequent hip fracture. (12)
Fragility fractures also have a significant impact on a person’s functional capacity and quality of life. A Canadian study reported that 40% of people were still unable to walk independently one year following a hip fracture. (13) Adachi et al. (14) reported a negative association between past osteoporotic fractures and health-related quality of life (HRQL) in both women and men that was dependent on fracture type and gender. For example, HRQL was significantly lower in both women and men who had experienced a hip fracture or a rib fracture compared with people without these fractures. The same study also found that women who had a past clinical vertebral deformity or a fracture in the lower limb had lower HRQL, largely because of pain, decreased mobility, and impaired ability for self-care, while a fracture in the lower limb was associated with decreased dexterity in men. Adachi et al. (14) reported that even a subclinical vertebral deformity in women was related to decreased cognition and increased pain, resulting in a lower HRQL.
Goeree et al. (2) estimated that osteoporosis and osteoporotic fractures in people 45 years and older cost the Canadian health system $1.3 billion (Cdn) in 1993, including $465 million in acute health care, $563 million in long-term care, and $279 million for chronic care hospitals. Based on the above estimation and population, Ontario’s proportion of the total osteoporosis cost would be approximately $400 million. The greatest portion of the cost was attributed to hip fractures. (2) In a 2001 Ontario study, the same investigators estimated the mean 1-year cost of a hip fracture in people aged 50 years and older to be $26,527 (Cdn) (95% confidence interval [CI], $24,564 – $28,490). The annual economic impact of hip fractures in Canada was expected to rise from $650 million (Cdn) at the time of the study to $2.4 billion (Cdn) by 2041. (12)
Because of the burden of illness from fractures, attempts are made to identify people at risk of osteoporosis and fractures, and intervene in order to reduce the risk of fractures.
Low bone mass has been found to be a major risk factor for fragility fractures. Bone is composed of an organic phase of mainly collagen I, an inorganic phase consisting mainly of calcium phosphate crystals, and a cellular component of osteoblasts and osteoclasts. Every year, the human body replaces 10% of its bone mass. Bone resorption occurs at the osteoclasts. Formation of new bone in the osteoblasts involves synthesis of the organic matrix, followed by deposition of calcium crystals, and a gradual maturation process, resulting in an increase in the amount and size of calcium crystals. After reaching peak bone mass at age 25 to 29 years, bone density begins to decline until age 65, and the rate of decline slightly decreases thereafter. (15;16)
Bone mineral density measurement has been the most common test used to screen for and diagnose osteoporosis. It measures the amount of calcium per unit area (grams/square cm) or per unit volume (grams/cubic cm) in the bone. Results of BMD tests are expressed as T-score and Z-score.
T-score is the number of standard deviations (SD) from the mean BMD for young (25–45 year olds) adults. A T-score of – 2.5 represents a BMD value that is 2.5 SD below the mean BMD for young adults.
Z-score is the number of SDs from the mean BMD value for people of the same age and gender.
T-scores and Z-scores vary according to the technique and reference populations.
Bone mass is a major determinant of bone strength. Laboratory studies have shown a high correlation between bone mineral content and the force needed to break a bone. The World Health Organization (WHO) had established criteria for the diagnosis of osteoporosis in postmenopausal Caucasian women based on BMD and the associated risks of fractures. (17) According to these criteria, a T-score of at least –1 is considered normal; a T-score between –1 and of –2.5 indicates below normal bone density, and osteoporosis is considered to be present if the T-score is less than –2.5 (Table 1).
It should be noted that BMD is a continuous value. Despite the WHO definitions, there is no established threshold or cut-off value of BMD to distinguish low-and high-risk people. The WHO committee did not have enough data to create definitions for men or other ethnic groups.
The assessment of BMD has been used for the selection of patients for osteoporosis treatment.
The WHO presently does not have diagnosis criteria for women less than 65 years of age or for men. Osteoporosis Canada is developing guidelines for BMD testing in men and in perimenopausal women (age 40 – 60 years).
In 2004, the International Society for Clinical Densitometry (ISCD) published its official position (18) that included diagnostic definitions for other populations in addition to the WHO classification (Table 2).
The ISCD Official Position states that the WHO classification should not be applied to healthy premenopausal women, and Z-scores rather than T-scores should be used. Furthermore, the document further states that the diagnosis of osteoporosis in premenopausal women should not be made on the basis of densitometric criteria alone. (18)
The Official position states that the WHO classification and T-scores should not be applied to children, and Z-scores should be used. The bone density may be described as low for chronological age if the Z-score is below –2.0. However, the ISCD cautions that the diagnosis of osteoporosis in children should not be made on the basis of densitometric criteria alone. (18)
Most studies on risk factors for fragility fractures were conducted in postmenopausal women.
The Canadian Multicentre Osteoporosis Study (CaMos) followed 5,143 postmenopausal Canadian women for 3 years and analyzed the association of potential risk factors for incident fractures. Papaioannou et al.(19) reported the following findings of this study:
Other primary studies support the following as risk factors for osteoporotic fractures in postmenopausal women:
Bone Mineral Density
A low baseline BMD has been found to be a strong predictor of fragility fractures. A T-score of-2 in the spine in women is associated with a 4- to 6-fold increase in the risk of new vertebral fractures, (20) and every SD below the mean femoral neck BMD for young adults increases the age-adjusted risk of hip fracture by 2.6 (21) However, classifying postmenopausal women as osteoporotic based on BMD alone only accounted for 18% of the documented osteoporotic fractures in the National Osteoporosis Risk Assessment, (22) suggesting that other risk factors need to be used to identify women with BMD above the osteoporotic threshold but who may be at high risk of fractures.
Age is another determinant of risk of fracture independent of BMD. Kanis et al. (23) showed that the same T-score is associated with a much higher risk of fracture at a more advanced age. For example a T-score of –2.5 at age 70 years is associated with a 10-year risk of osteoporotic fracture of 24% compared with 12% at age 50 years with the same T-score.
History of Previous Fractures
This risk factor will be discussed in greater detail in the literature review section.
Family history of fractures
Increased risk of fracture has been reported for people with a maternal history of fractures, particularly hip fractures. This effect appears to be most pronounced on future hip fractures. For example, Taylor et al. (24) reported in the Study of Osteoporotic Fractures that in elderly white women, a history of maternal hip fracture increased the risk for subsequent hip fracture independent of BMD (hazard ratio [HR] adjusted for BMD 1.35, 95% CI, 1.14–1.5]). Albrand et al. (25) reported in the Os des Femmes de Lyon (OFELY) cohort study that the odds ratio (OR) for 5-year risk of fragility fractures was 1.77 (95% CI, 1.01–3.09, P = .04) in healthy postmenopausal women who had a maternal history of fragility fractures. However, Bensen et al. (26) found that maternal history of fracture was a significant predictor of future fractures only at the rib (OR, 2.89; 95% CI, 1.035–8.081).
Propensity to falls
A tendency to fall has been identified as a predominant nonskeletal predictor of fragility fractures in the elderly. (27) It has been reported that about 90% of hip fractures involve falls. (28) Kaptoge et al. (29) found in the prospective multinational Europian Prospective Osteoporosis Study (EPOS) that BMD appeared to be less important in explaining variations in incidence of upper limb fractures in women across diverse populations in Europe, compared with the effect of location-specific risks of falling and factors that may be associated with the likelihood of falling. The nature of the fall likely determines the type of fracture, while bone density and factors that increase or attenuate the force of impact of the fall determine whether a fracture will occur when a faller lands on a particular bone. (28) The majority of falls in old age likely result from a combination of factors relating to aging and poor health, such as decreases in muscle strength and function, gait disorders, and loss of balance. (30) Epilepsy, use of seizure medication, Parkinson’s disease, and wearing corrective lenses are factors that tend to be associated with increased risk of pelvis fracture in men and women. (31)
Other risk factors
The reduction in estrogen associated with menopause was found to be the strongest risk factor for osteoporosis in women. Other risk factors for osteoporosis are low body weight (body mass index [BMI]<20 kg/cm2), lack of weight bearing activity, cigarette smoking, low dietary calcium/vitamin D, certain medications (e.g. corticosteroids, chronic anticonvulsant therapy), and some health conditions (e.g. malabsorption syndrome and primary hyperparathyroidism).
All the above risk factors should be used in conjunction with bone mineral measurements to assess an individual’s overall risk of fragility fractures and need for treatment.
Reducing the prevalence of osteoporosis requires preventative life style changes such as increasing dietary intake of vitamin D and calcium, increasing physical activity, smoking cessation, and moderating alcohol intake. Current available treatments for osteoporosis are mainly medications that reduce bone resorption (hormone replacement, bisphosphonates, estrogen receptor modulators, salmon calcitonin, and parathyroid hormone). The effectiveness of these treatments will be discussed in greater detail later.
Because of the silent nature of osteoporosis, early detection and treatment to prevent fractures in people with low BMD has been recommended. As clinical risk factors do not have adequate accuracy to identify patients with osteoporosis, BMD measurements are used in the setting of clinical risk factors for fractures to assess whether there is also low BMD that further increases the risk for fractures. (15)
There are many techniques for measuring bone mineral density. They fall into 2 main categories: those that use ionizing radiation and those that do not. Ionizing techniques include:
BMD measurement of the entire skeleton or specific sites such as the spine or hip using x-ray absorptiometry is based on the absorption of x-rays by the calcium crystals in the bone. A dual energy x-ray absorptiometry (DXA) machine sends a thin beam of low-dose x-rays with 2 distinct energy peaks through the bones of patients. One peak is absorbed mainly by soft tissue and the other by bone. BMD is calculated from the difference in absorption between the bone and the soft tissue. Computer software calculates the numerical density of the bone from the image and compares it with the mean of healthy young adults, and to the age-matched control of the reference population. A radiologist interprets the data and creates a concise report on the patient’s bone density status. The BMD test with DXA takes approximately 10 to 30 minutes, and the dose of radiation received by the patient is equivalent to one-fifth to one-half of the dose from a chest x-ray. The accuracy of DXA at the hip exceeds 90%. New developments in DXA include the use of multi-element detector array with true fan-beam, single sweep scanning, and concomitant lateral vertebral assessment to screen for vertebral fractures. Small DXA devices are also available for measuring BMD in the heel or forearm in as little as 15 seconds. The distal radius is often used because it contains trabecular and cortical bone. Presence of osteomalacia or osteoarthritis may result in a high BMD value that does not reflect higher bone strength. (32)
Single X-ray absorptiometry (SXA) is similar to DXA but uses a single beam to measure BMD of the wrist or the heel.
Quantitative computed tomography (QCT) can be performed on the spine using standard CT devices. QCT assesses 3-dimensional bone density and permits isolated measurement of trabecular bone density; however, QCT is not widely used because its reproducibility is poor and it exposes patients to far too high a radiation dose to be acceptable. The clinical utility of smaller peripheral QCT devices is also being investigated.
Other ionizing radiation techniques include radiograph of proximal phalanges, single photon absorptiometry (SPA) and double photon absorptiometry (DPA). These methods are no longer in use for BMD measurements. (personal communication, clinical expert, August 2006)
Techniques that do not use ionizing radiation are:
The transmission of sound through bone reflects its density and structure and can be assessed quantitatively using the speed of sound or broadband ultrasound attenuation. Quantitative ultrasonography (QUS) of the heel resembles other peripheral measurements in terms of ability to predict fractures. QUS is noninvasive, involves no exposure to ionizing radiation, and is less expensive and portable. However, there is a need for normative data, quality assurance programs, standardization, and attention to precision, sensitivity, and accuracy. (33) Experts advised that it has not been widely used because of low precision. (Personal communications October 2006)
Measurement of biochemical bone markers in the blood may provide information on bone remodelling. Bone markers are specific for bone formation (e.g. bone alkaline phosphatase) or bone resorption (e.g. deoxypyridinoline), and may be influenced by age, gender, ethnicity, menopause status, diseases, recent fractures, immobility, treatment, and timing of sample collection.(34) Bone markers cannot be used to diagnose osteoporosis. However, studies suggest that bone markers used in conjunction with BMD may improve the prediction of fracture risk. (35;36) Since bone marker levels change quickly with the initiation of osteoporosis treatment, they may be used as a surrogate marker for treatment efficacy. The use of bone markers is limited by its variability.
Since microstructure is a determinant of bone strength, techniques are being investigated for measuring this parameter. In recent years, methods of high-resolution magnetic resonance imaging (MRI) are being investigated for the assessment of bone density and its microstructure. (37;38)
A new portable device called a mechanical response tissue analyzer is being investigated as a means of measuring the mechanical properties of the ulna and tibia to reflect both mineral content, and geometry/ structure of the bone. This device is not yet available for clinical use.
Bone mineral density measurement yields different results depending on the technique used and the site of measurement. Correlation between results from different techniques has been poor. Presently, BMD measured with DXA at the hip and/or spine is considered the gold standard for the noninvasive diagnosis of osteoporosis, and has been used by the WHO to define osteoporosis. (17) Diagnosis is based on the lowest BMD obtained. The ISCD (18) recommends that BMD should be measured at both posterior-anterior spine (L1–L4) and hip (proximal femur, femoral neck, or trochanter) in all patients, and forearm (33% radius) BMD should be measured when the hip or spine cannot be measured or interpreted or in cases of hyperparathyroidism or very obese patients. The ISCD indicated that spine BMD should be interpreted with caution in the elderly because degenerative arthritis in the posterior elements of the spine may result in an artifactual increase in measured BMD. Furthermore, the ISCD also stated that the WHO classification for the diagnosis of osteoporosis and osteopenia should not be used with peripheral BMD measurements other than BMD at 33% radius. (18)
Based on the WHO definitions of osteoporosis, the T-scores and Z-scores will vary depending on the reference standards used. In Canada and the United States, the densitometers are programmed to use normative data from the United States National Health and Nutrition Examination Survey (NHANES III, a population-based study), for white, black, and Asian subgroups, and for men and women. (16;39;40) Peak bone mass for the Canadian population has been established in the Canadian Multicenter Osteoporosis Study (CaMos) using 10,061 women and men aged 25 years or more randomly selected from 9 regions across Canada. (3) However, this database has not been used routinely in the interpretation of BMD tests. There is controversy over whether thresholds derived from women can be applied to men. Studies have shown that for hip and vertebral fractures, the 10-year risks of fracture are similar in men and women for T-scores close to the diagnostic thresholds, lending support that T-scores derived from women are applicable to men, and that diagnostic thresholds should be the same in men and women. (41)
Parameters usually included in current BMD reports are shown in Appendix 1. In the Recommendations for Bone Mineral Density Reporting in Canada, the Canadian Association of Radiologists recommended including fracture risk (low, moderate, or high) in BMD reports, stratified by gender and age group, and T-score, as well as a patient questionnaire that identifies the patient’s clinical risk factors. This recommendation aims to integrate BMD measurements with other clinical factors to quantify fracture risk assessment. (42)
Under the Ontario Osteoporosis Strategy, Osteoporosis Canada is submitting guidelines for BMD testing and reporting.
Health Canada has licensed numerous bone mineral densitometers as class 3 devices. These are summarized in Appendix 2.
Many health agencies and jurisdictional governments have developed recommendations for BMD testing. These recommendations are summarized in Appendix 3. None recommended using BMD to screen for osteoporosis in the general population.
Almost all of these recommendations target women over 65 years, and for women younger than 65 years of age, only those with risk factors for osteoporosis are targeted. The British Columbia Guidelines for Bone Density Measurement in Women emphasized that even in the presence of risk factors, BMD measurement should only performed when the results are likely to alter patient care. (43) The Canadian Task Force for Preventive Care recommended the use of risk assessment instruments for case finding, to further identify people at risk who should undergo BMD testing.(44)
Only Osteoporosis Canada (45) and the ISCD made recommendations for men (See section on BMD Testing in Men).
Conditions for BMD Tests in the Ontario Health Insurance Plan (OHIP)
In the OHIP Schedule of Benefits (46), only bone mineral testing by axial technique using DXA at the hip/and or spine is an insured service. The conditions for the service to be insured include:
For the purpose of this service, “high-risk patients” means a patient at risk of accelerated bone loss due to either states of high bone turnover such as primary hyperthyroidism and glucocorticoid induced osteopenia, or due to such other conditions as have been determined by the Scientific Advisory Board of the Osteoporotic Society of Canada OSC (presently Osteoporosis Canada) which prevail at the time the service is rendered. “Low-risk patient” means any patient who is not a high-risk patient (Table 3).
The OSC guidelines (47) recommend BMD testing for everyone over the age of 65 years and also for people over age 50 if they have at least one major or two minor risk factors. According to the OSC guidelines:
Major risk factors for osteoporosis include:
Minor risk factors for osteoporosis include:
The main purpose of BMD tests is to identify people with osteoporosis and treat them effectively in order to reduce their risk of fragility fractures. Several studies had been published on the use of BMD tests in Ontario. The most recent publication by Jaglal et al. (48) reported that BMD tests in the province increased 10-fold between 1992 and 2001. The same study also suggested that the increase in BMD use was accompanied by an increase in the use of bone sparing medication, and a decrease in the rate of hip fractures and wrist fractures in people aged 65 years or more.
There are 2 main parts in this assessment:
Part 1 - Objectives-
Part 2 - Objectives
Analyses of administrative data were conducted to update the trend of BMD test utilization in Ontario, and the corresponding trends in the incidence of fragility fractures in the province.
Questions to be addressed by the analyses:
The Institute of Clinical Evaluative Sciences (ICES), under the direction of Susan Jaglal, Ph.D., abstracted data from administrative databases. For all data, only Ontario residents were included.
The codes and criteria for data abstraction are shown in Appendix 4. The data were analyzed to identify trends in BMD use from 1992 to 2005.
The number of BMD claims increased almost10-fold between 1993 and 2005, reaching 500,000 tests in 2005. Figure 1 shows that the largest increase in BMD testing occurred in 1996/97 (64%) and 1997/98 (56%). A change to OHIP coverage was implemented in October 1999, restricting reimbursement to once in any 24-month period for people at low risk. The use of BMD in 2000/01 increased by 12% compared with an increase of 20% in the previous year. The use of BMD actually decreased slightly in 2003/2004, probably due to restricted access to hospitals during the SARS epidemics in Toronto. In the last 2 fiscal years (2004 & 2005), the volume of BMD tests increased at a rate of 6 to 7% per year (an average increase of 30,000 tests per year). Approximately 90% of all BMD tests were performed in women (Figure 2).
The rate of BMD use per 100,000 population was examined to determine the increase in BMD use independent of population growth. Figure 3 shows that with the exception of 2003/04, the rate of BMD tests has increased steadily and is still increasing in the last 2 fiscal years at a rate of 5% to 6% per year. It was estimated that more than 80% of BMD tests were ordered by family practice physicians.
Since women accounted for 90% of all BMD tests, the data pertaining to women were further analyzed. Figure 4 shows that the rate of BMD use was highest in women aged 65 to 69 years, followed by the 70 to 74 and 75 to 79 age groups. The rate was about the same for women under age 40 to 64 years and women aged 80 to 84 years. Women older than age 85 years had the lowest rate of BMD tests. In the most recent 4 fiscal years, the increase in the rate of BMD use occurred in the greater than 65 year age groups, while the rate for the 45 to 65 year age group remained more or less constant.
Even though women under age 65 years had the second lowest rate of BMD tests, due to the number of women in this age group, they accounted for 61% of all BMD tests. The highest number of BMD performed was in women in the 55 to 59 and 50 to 55 age groups (Figure 5).
The rate of BMD tests varied across geographical areas. The rates were highest in the Central, Mississauga Halton, Toronto Central, Hamilton Niagara Haldimand Brant, and Central East local health integrated networks (LHINs). LHINs in remote areas had the lowest rates of BMD testing (Erie St. Clair, North East and North West). The greatest geographical variation in rates of BMD testing was between the Central LHIN and the North West LHIN. However, trend analysis shows that this gap is slowly decreasing over time. In 2002, the rate was almost 3-fold higher in the Central LHIN compared with the North West LHIN. In 2003, this gap has decreased to 2.2-fold (Table 4).
Jaglal et al. (49) previously reported a 17-fold variation in the BMD rate across counties in 2000. Although the gap in BMD rates across counties had narrowed by 2004, there was still an almost 10-fold variation between the highest rate (203.76 per 1,000 women for Toronto) and the lowest rate (21.04 per 1,000 women for Kenora) (Table 5). The overall rate was 142 per 1,000 women. Most of the more remote or rural counties had lower BMD rates.
Bone mineral density tests performed in 2003/04, 2004/05, and 2005/06 were analyzed according to the history of BMD tests in the previous 2 years. Four patterns of BMD testing were observed: category 1 includes patients who had a BMD test in the previous year (annually for 2 years); in category 2, a BMD test was performed 2 years prior to the current tests (a repeat test in a 24 month period); category 3 represents people with a BMD test annually 3 years in a row; and category 4 represents people with no BMD in the previous 2 years (Table 6). The distribution of BMD tests among the 4 categories was quite consistent for the 3 fiscal years studied. The data showed that approximately 59% of people tested did not have a BMD test in the previous 2 years, while 41% had a repeat test during a 3-year period. About 17 to 18% had a BMD annually within a 2-year or 3-year period, and 23% to 24% had a repeat BMD test in 2-year period.
The profiles of patients in each of the categories were analyzed according to age (<65 vs ≥65 years), presence or absence of fracture in the most recent year, and the risk for osteoporosis according to the fee code.
For people who had a BMD test in 2005/06 but no BMD in the previous 2 years (category 4), 34% (>98,000 BMD tests) were performed in people under age 65 years, who had no fracture in the year of study, and were also coded as low risk in the OHIP claims (Table 7). Although the risk level of these patients cannot be validated and some might have been miscoded as low risk, even if 50% were coded correctly, it would mean 49,000 BMD tests were performed in people less than 65 years of age and at low risk of osteoporosis and fractures, which is not consistent with current Canadian guidelines for BMD testing.
More than 24,000 BMD tests performed in 2005/06 were repeat BMDs. Of the people who had annual BMD tests for 2 or 3 years (categories 1 or 3), more than 90% were considered high risk (Table 7) and were in compliance with OHIP since there are no restrictions on BMD testing in high-risk people. However, approximately 3,500 annual repeats were performed in people rated as low risk, in contravention to the OHIP conditions, since people at low risk are limited to one BMD in any 24 month period.
In each of the 3 years studied, approximately 21,000 repeat BMD tests within a 24 month period were performed in people under 65 years of age, were coded as low risk, and had no fracture during the year of the study (Table 7). Although these repeat tests were compliant with OHIP reimbursement policies for BMD, these policies were last revised in 1999. The evidence for serial BMD tests every 2 years in people at low risk of osteoporosis needs to be re-examined.
The percentages of people (age > 65 years) who underwent BMD testing and/or had a ODB claim for osteoporosis drugs during the first year after a fragility fracture are summarized in Table 8, and 1 year after a hip fracture in Table 9. For people who had no BMD during the first year after a fracture, the database was searched to verify that no BMD tests were performed in the 5 years prior to the fracture.
The analysis showed that only approximately 19% of people who had a fragility fracture in 2003 or 2004 had a BMD test within the first year following the fracture or in the previous 5-year period, and approximately 40% received antiresorptive pharmacologic treatment during the 1-year period. This means that 81% of patients age 65 years and older did not have bone density assessment following a fragility fracture, representing approximately 9,100 people in 2004/2005. Patients were more likely to receive antiresorptive treatment if they had a BMD assessment after fracture compared with those who did not (64% vs 35%) (Table 8). Approximately 3,200 people had ODB prescription for antiresorptive drugs in the year after a hip or wrist fracture without any baseline BMD measurements (after fracture or in the previous 5 years).
The percentage of people who had a BMD test during the first year following a hip fracture was even lower (approximately 13%) while the percentage that received antiresorptive treatment in the same period was 45%. Approximately 4,500 patients who had a hip fracture did not undergo a BMD assessment 1 year after the fracture. The rate of treatment was higher in patients who had a BMD test compared with patients with no BMD test after the hip fracture (75% vs 41%) (Table 9).
Due to greater bone mass, osteoporotic fractures tend to occur later in men compared with women. Osteoporosis is being recognized as a growing concern in men because of the increasing average life expectancy in men. Although the prevalence of osteoporosis (as defined by WHO) in Canadian men 50 years of age and over is 5% compared with 16% in women of the same age, osteoporotic fracture in men results in higher mortality rates than in women (Appendix 5). Standardized mortality ratios for people who have had major fractures compared with the general population ranges from 2.3 to 3.2 for men, and from 1.7 to 2.2 for women. (50) Osteoporotic fractures in men also results in increased morbidity and reduced quality of life due to decreases in mobility, independent living, and dexterity.(48) Moreover, radiographic studies show that the prevalence of vertebral deformity (fracture) is about 25% in both genders.
Current utilization data suggests that BMD tests have been underutilized as a screening tool in men. Even though men accounted for about 24% of hip fractures and 21% of wrist fractures in people aged 50 years and older (2004 & 2005), only 10% of all BMDs were performed in men. Men were also less likely than women to undergo BMD measurements (13 % vs 21%) or receive antiresorptive treatment (20% vs 46%) following a fracture (Table 10).
Figure 6 shows that the number of Ontarians aged 65 years and older who filled prescriptions for antiresorptive medications has increased steadily since 1996 The rate of increase in the number of people was 12.2% and 9.2% for 2004/05 and 2005/06 respectively, reaching a peak of more than 280,000 people, and 1.626 million prescriptions in 2005/06. The type and percentage of the different antiresorptive drugs prescribed for people receiving ODB coverage are summarized in Table 11.
The rate of both wrist fractures and hip fractures in women started to decline in 1996 (Figure 9). This decline continued for hip fractures, reaching a low of 24.2 per 10,000 women (41% reduction since 1992). The rate of wrist fractures in women plateaued in 2003 and started to rise again in 2004, and reached a rate of 44.9 per 10,000. The rates of wrist fractures and hip fractures had also declined in men with a larger decrease in the rate of hip fractures (Figure 7).
The total number of hip fractures has remained relatively stable in the last 4 years but the number of wrist fractures started to rise again in 2003 (Figure 8).
Based on the above analysis, a steady increase in the use of antiresorptive drugs and a decrease in the rate of hip and wrist fracture (particularly hip fracture) occurred in the same period (1997 - present), during which BMD testing escalated.
The rate of hip fractures was low for people under 65 years of age (Figures 9 and and10).10). For women, the rate began to increase in the age group of 60 to 65 years, and continued to increase exponentially with the sharpest increase occurring after age 80 years (Figure 9). The same pattern occurred in men but the increase in rate appeared to occur about 5 years later than in women (Figure 10).
There seems to be a small increase in the rate of wrist fractures around age 55 to 60 years in women, with the highest rise in rate occurring at age 70 years (Figure 11). For men, the rate of wrist fractures remains low until age 70 years (Figure 10).
Even though there appears to be room for improvement in the utilization of BMD testing, the increase in BMD use in Ontario appears to be having a positive impact on the rates of hip and wrist fractures, Future strategies should not focus on a blanket reduction in the use of BMD assessment. Instead, they need to ensure that the test is being performed in people at high risk of osteoporosis and fractures. A literature review was conducted on issues relating to the overuse and gaps in BMD testing to inform future polices relating to BMD service in the province.
Separate search strategies were developed to address the main questions analyzed in the systematic review. The detailed search strategies are shown in Appendix 6. All searches were run between May 5 and August 30th, 2006 in the following databases: OVID MEDLINE, OVID MEDLINE In-Process & Other Non-indexed Citations, OVID EMBASE, Cochrane Library, and the INAHTA/CRD database. All searches were limited to human subjects and English-language articles. Additional searches of websites and references of publications were also performed to ensure comprehensiveness.
The first search strategy (detailed in Appendix 6a) was developed to locate published articles that evaluated the relationship between changes in BMD (as a result of pharmacologic therapies for osteoporosis), and fracture and fracture risk. The pharmacological therapies of interest included etidronate, alendronate, risedronate, raloxifene, estrogen replacement, parathyroid hormone, and calcitonin. This search was limited to articles published between January 2000 and August 29, 2006, and yielded 297 citations.
The second search strategy (detailed in Appendix 6b) was developed to locate published articles dealing with BMD testing and treatment rates for osteoporosis after a fragility fracture. This search was limited to articles published between January 2005 and August 30, 2006, as a very comprehensive systematic review (51) was published in 2005 which included studies published through 2004. This search yielded 333 citations.
The third search strategy (detailed in Appendix 6c) was developed to identify randomized controlled trials (RCTs) of specific pharmacological treatments for osteoporosis. This search was limited to meta-analyses, systematic reviews, and randomized controlled trials published from January 2001 to June 19, 2006, and yielded 458 citations. As an addendum to this search, major RCTS published between 1997 and 2001 were included to ensure comprehensiveness
The fourth search strategy (detailed in Appendix 6d) examined the predictive value of BMD in men. The search was limited to articles published between January 1, 2001 and May 4, 2006, and yielded 381 citations.
General Inclusion Criteria
This review included published English-language journal articles that reported primary epidemiological or clinical data if:
For rate of change of bone loss for women and men not receiving osteoporosis treatment
For effectiveness of BMD tests in monitoring response to osteoporosis treatment
For rate of change in BMD during treatment for osteoporosis
For patient compliance with osteoporosis therapy
For prevalence of BMD tests after a fragility fracture
For impact of a previous fracture on the risk of future fractures in men and women
For effectiveness of osteoporosis treatment in reducing risk of fractures
For impact of BMD test on treatment
For predictors of fractures in men
Studies were excluded if they:
One researcher selected reports based on inclusion and exclusion criteria. In total, 28 systematic reviews/meta-analyses and 97 primary studies (34 randomized controlled trials and 63 observational studies) were included.
One researcher reviewed the full-text reports and extracted data using data extraction forms. For RCTs, the quality of studies was assessed using criteria adapted from Jadad et al. (52) The quality of observational studies was evaluated based on method of patient selection, sample size, statistical analysis and completeness of follow-up. Levels of evidence were assigned to studies included in each section (Tables 12A to 12I).
Eight meta-analyses of cohort studies and 28 primary studies met the selection criteria. The studies are classified in Table 7A.
Analysis of the administrative data showed that more than 18% (21,000) of all BMD tests in 2005/06 were repeat tests performed within 24 months in people at low risk of developing osteoporosis, in compliance with current OHIP reimbursement policies. The importance of a baseline BMD in assessing the risk of fractures has been well established; however, the utility and frequency of serial BMD measurements are less well defined. For this reason, evidence relating to the utility and frequency of serial BMD screening and serial BMD monitoring during osteoporosis treatment were examined.
For people who are not receiving osteoporosis treatment, the purpose of repeating BMD measurements is to monitor the progression of bone loss in order to initiate treatment when necessary. The time interval from baseline test for the BMD to drop to a treatment level depends on 4 parameters:
The last parameter would depend on the presence or absence of other risk factors for fractures.
The reliability of a follow-up BMD test depends on the precision (reproducibility) of the specific test, i.e., the ability of the test to produce the same results in repeated measurements of the same individual. Factors that can affect the precision of a BMD test are equipment, operator, patient population, site of measurement, and positioning of the patient. (53)
Precision can be stated either as the SD, or as the percent coefficient of variation (%CV), defined as %SD/mean. Percent CV of DXA tests have been reported to range from 1.8% to 2.3% at the lumbar spine, 2.3% to 3.6% at the femoral neck, and 1.7% to 2.5% for the total hip. (54) Precision may vary widely among DXA facilities. In a 7-centre pharmaceutical trial, 6 of the sites showed BMD test precisions at the posteroanterior spine ranging from 0.969 to 2.101%, and at the femoral neck ranging from 1.475 to 3.362%. However, the 7th site yielded an average PA spine precision of 3.565%, and a femoral neck precision of 4.349%. A change in BMD between the baseline and a repeat test may reflect the precision (reproducibility) of the test rather than a real biological change in bone density. Hence, interpretation of serial measurements can only be accomplished with the knowledge of the precision of the specific DXA facility where the test was performed. (53)
Changes in 2 BMD measurements at the same skeletal site in an individual may be related to measurement errors unless they are greater than the least significant change (LSC) (LSC at 95% CI = 2.77 × % CV) or the smallest detectable difference (SDD = 2 × SD in g/cm2). (55) Maghraoui et al. (56) recently reported at their center an LSC of 3.56% (total hip) and 5.60% (spine) and an SDD of 0.02g/cm2 (total hip) and 0.04 g/cm2(spine). There are indications that absolute precision errors derived from the SD are preferred because they are independent of the level of BMD. Precision errors and LSCs of a BMD facility are usually measured by performing 2 or more scans on a group of patients and then calculating the root-mean squared standard deviation of the replicate measurements. The ISCD recommends either measuring 30 subjects twice, or 15 subjects 3 times.(53)
Table 13 illustrates the impact of precision on the interval of serial BMD measurements. For a BMD facility with a 1% precision, the LSC is close to 3%. At a rate of BMD change of 1% per year, the shortest time interval to repeat the BMD measurement to obtain a result that exceeds the least significant change is 3 years. Similarly, at the same rate of BMD change (1%), it would take 6 years for the BMD change to exceed the LSC (6%) for a BMD test that has a 2% precision. Conversely, in situations where a rapid rate of bone change is expected (e.g., in certain disease conditions or during treatment), the LSC may be exceeded within a shorter period of time (e.g., 1 year for a precision of 1% and a rate of change of 3%/year).
In order to reduce the random error, repeat BMD measurements should be made at the same facility using the same instrument and same scanning procedure. The Canadian standards and guidelines for DXA densitometry recommend a documented quality control program at each DXA facility to ensure minimal radiation control, proper calibration, and ongoing monitoring of precision. (54)
A baseline BMD has been shown to predict 10-year risks of fractures. Initiation of osteoporosis treatment is usually made on the basis of risk of fracture and BMD values. The closer the baseline BMD is closer to the osteoporotic range, the sooner treatment may be needed, and hence more frequent monitoring of BMD will be necessary.
The rate of bone loss influences how quickly the BMD may reach a level that requires treatment. This in turn, will determine the interval between the baseline BMD and repeat BMD measurement.
A systematic review was conducted to determine the rate of BMD changes in women and men who were not receiving osteoporotic treatment. The review included 1 prospective population-based study in men and women, 5 prospective population-based cohort studies in women (sample ranging from 50 to 1,035), and 6 cohort studies in men (sample ranging from 214 to 5,995. Rates of bone loss from baseline in the placebo arm of RCTs on osteoporosis treatment were also reviewed, including 10 RCTs for postmenopausal women, and 2 RCTs for men with primary osteoporosis. These studies are described in Appendix 7, and mean rates of BMD changes are summarized in Appendices 8 and 9. The data suggest that in postmenopausal women, the mean rate of BMD loss at the spine, femoral neck, or total hip is generally 1% per year or less. Chapurlat (57) showed that apart from a small but significant bone loss at the hip, trochanter and anteriorposterior spine in premenopausal women (< 0.1% to 0.3% per year), there was no significant bone loss at other skeletal sites. Even though perimenopausal women showed more rapid bone loss compared with premenopausal women, the mean rate of loss is still less than 1% per year at all skeletal sites. (57) Even in the placebo arm of osteoporotic treatment studies, postmenopausal women with osteopenia or osteoporosis had a mean rate of bone loss of 1% per year or less. It should be noted that even though the mean rate of bone loss is up to 1% per year, there were individual women who had lost more than 1% of BMD per year. In women who have other risk factors such as lactose intolerance or surgical menopause (without hormone replacement), the rate of bone loss can be as high as 3% per year. (58;59)
Appendices 9A and 9B show that the mean rate of bone loss in men who do not have other major risk factors is also generally less than 1% per year; however, there were exceptions. Ensrud et al. (60) reported that men who had lost at least 5% of their body weight had a significant increase in the rate of bone loss compared with people with stable weight or weight gain (mean –1.2% to –1.7%, range –0.9% to –2.%) (Appendix 9B). The increased rate of bone loss occurred regardless of baseline BMI or whether the weight loss was voluntary. A similar relationship between weight loss and increased bone loss was observed in women. (61)
The higher the rate of bone loss, the sooner BMD may progress to a level at which treatment becomes necessary, and hence more frequent monitoring may be required.
Apart from baseline BMD and the rate of bone loss, presence of risk factors for fractures will also influence the BMD threshold at which treatment needs to be initiated and hence the frequency of BMD monitoring.
As previously discussed, the most important predictors of fractures independent of BMD are:
Experts advised that for people with any of the above risk factors, treatment may be considered at a higher BMD level (e.g. at a T-score of –1.5 to –2), whereas for people who did not have any of these risk factors, treatment may not be necessary until a T-score of –2 to –2.5).
The progression of BMD over time was computed for 3 scenarios using baseline T-scores of 0 (peak BMD), –1 (lower limit of normal) and –1.5 (osteopenic), and using a rate of change of BMD loss of 1% to 5% in each scenario (Tables 14A to 14 C). It was assumed that pharmacologic treatment would begin at a T-score of –2 (before the patient becomes osteoporotic). For each scenario, the time intervals for the BMD loss to exceed the LSC for test precisions of 1% to 3% were computed. The time intervals to reach the treatment threshold were also identified.
The model showed that regardless of the baseline BMD, for a normal rate of bone loss of 1% per year (as in the majority of people), it will take approximately 3 years for the BMD loss to exceed the LSC of a BMD test with a 1% precision (Tables 14A – 14C). This means that for the majority of people, even with the most precise BMD test, repeating the test less than 3 years after the baseline test is not likely to provide meaningful information. The less precise the test, the longer it would take to detect a BMD change that can be considered significant (e.g. 9 years for a test with a 3% precision error). It is, therefore, important that BMD facilities have high precision so that genuine biological bone loss in people who require frequent monitoring can be detected Conversely, as the rate of bone loss increases, it will take less time to exceed the LSC and for a genuine change in BMD to be detected. For example, at a rate of bone loss of 1% per year, the LSC for a 1% precision will be exceeded in 3 years whereas with a bone loss of 3% per year, the LSC for a test with the same precision will be exceeded in 1 year (Tables 14A–C).
Table 14A shows that for a person with a normal baseline T-score of 0 and a normal rate of bone loss up to 1% per year, the BMD probably will not drop to an osteoporotic level within 25 years. Even at an increased rate of bone loss of 3%, it would take 13 years to reach a T-score of –2 and 10 years to reach a T-score of –2.5, levels at which treatment is usually considered.
At a baseline T-score of –1 (lower limit of the normal BMD range) and at a normal rate of bone loss (up to 1% per year), it would take approximately 16 years for the BMD to drop to a T-score of –2, the level treatment may need to be initiate (Table 14B). Repeating the BMD test every 2 years after the baseline BMD test in this population will not serve any clinical purpose.
However, if a person with a normal baseline T-score of –1 has increased rate of bone loss of 3% per year, the time interval for the BMD to drop to the treatment level (T-score of –2) will be shorten to 5 years (Table 14B). If the person is osteopenic at baseline (e.g. with a T-score of –1.5), and is losing BMD at an accelerated rate of 3% per year, the BMD may drop to a T-score of -2 within 3 years (Table 14C). The lower the baseline BMD and the higher the rate of bone loss, the sooner treatment may need to be initiated, and the closer the BMD needs to be monitored.
Thus, for people with a normal baseline BMD and a normal rate of bone loss without major risk of fractures, BMD measurements do not need to be repeated every 2 years. A repeat BMD measurement after the baseline test will be required (in 3–5 years depending on the precision of the test facility) to establish the rate of bone loss. The frequency of further serial testing should be guided by the baseline BMD and the rate of bone loss (see Tables 14A to 14C). For people that have osteopenic baseline BMD, have accelerated rate of bone loss, or have major predisposing factors for osteoporosis, BMD needs to be monitored more frequently so that treatment can be initiated promptly (Table 15).
For men and women receiving glucocorticoid treatment, BMD loss can be as high as 8% in 20 weeks. (62) Hence, for patients receiving 7.5 mg corticosteroid per day or more for at least 3 months, BMD monitoring may be performed every 6 to 12 months depending whether the person is receiving osteoporosis treatment.
|*Major Risk Factors|
|-||History of a fragility fracture and age greater than 75 years|
|-||Fragility fracture in first degree relative (particularly maternal history of hip fracture)|
|-||A condition associated with rapid bone loss|
|-||Medication associated with rapid bone loss|
In order to determine whether BMD is useful for monitoring of response to treatment, the following questions must be answered:
To be considered an appropriate surrogate for fracture as a treatment endpoint, BMD should be related causally to a decreased propensity to fracture, and the change in BMD must largely capture the intervention’s effect on the propensity to fracture. (63) Seven meta-analyses and 2 RCTs on this subject were identified (Table 16).
The quality assessment and description of these meta-analyses is summarized in Appendices 10, 11 and 12.
Two meta-analyses reported that an increase in BMD is responsible for most of the reduction in fracture risk.
Wasnich et al. (64) conducted a meta-analysis of 13 randomized, placebo-controlled trials of antiresorptive drugs including etidronate, alendronate, tiludronate, raloxifene, hormone replacement therapy (HRT), and calcitonin with a total follow-up of 63,822 person years. Overall, trials reported that patients with a larger increase in BMD tended to have greater reductions in vertebral fracture risk. Poisson regression showed that treatments that increase spine BMD by 8% reduced vertebral fracture risk by 54%, and changes in BMD explained most of the total treatment effect on fracture risk (41% risk reduction). A risk reduction of 20 to 22% was not associated with any measurable change in spine BMD.
Hochberg et al. (66) also found a significant association between the amount of increase in BMD at the spine and hip and nonvertebral fracture risk reduction. Hochberg et al. (66) conducted a meta-analysis using trial level summary data from 18 randomized, double-blind, and placebo-controlled trials on antiresorptive drugs including etidronate, alendronate, tiludronate, risedronate, raloxifene, estradiol, and calcitonin, with a total sample size of 26,494 women with incident nonvertebral fractures and 69,369 women-years. The analyses showed that larger increases in BMD at both the lumbar spine and hip during treatment were significantly associated with a greater reduction in the risk of nonvertebral fracture (P= .02 and .06 respectively). Each 1% increase in spine BMD at 1 year was associated with an 8% reduction in nonvertebral fracture risk (P= .02) An antiresorptive drug that increased spine BMD by 6% at 1 year reduced nonvertebral fracture risk by about 39%, and one that increased hip BMD by 3% at 1 year reduced nonvertebral fracture risk by about 46%. Hochberg et al. concluded that changes in BMD appeared to explain a significant part of the fracture risk reduction.
There is growing evidence that an increase in BMD alone does not fully account for the reduction in fracture risks.
Guyatt et al. (63) summarized the results of 8 meta-analyses that examined the magnitude of effects of osteoporosis therapies on fracture and bone density and conducted an analysis of the relationship between changes in bone density and magnitude of fracture reduction. Based on a regression analysis using data from systematic reviews and results from a large RCT of parathyroid hormone, Guyatt et al. (63) found a 20% reduction in the relative risk for vertebral fracture which was not associated with changes in BMD, and an additional 25% reduction in relative risk that was associated with changes in BMD. Based on the analysis, Guyatt et al. concluded that BMD is not helpful for predicting the impact of antiresorptive treatment on nonvertebral fractures.
Dalmas et al. (67) repeated the meta-analysis by Hochberg et al., (66) using individual patient data from all but 3 of the same studies. The results showed that there was no association between the extent of a reduction in nonvertebral fracture risk and an increase in BMD at the spine or hip at 1 year or at study end point. Larger increases in the lumbar spine BMD at 1 year were not associated with a greater reduction in nonvertebral fracture risk (P= .12), nor was there any association with increases in hip BMD (PP= .11).
Even when there was a significant increase in BMD, it has not been shown that BMD gain is a determinant of treatment effectiveness. In a meta-analysis by Cummings et al., (65) using individual patient data of 12 blinded, randomized, placebo-controlled trials of antiresorptive drugs, a 1% increase in spine BMD in the treatment group (compared with placebo group) was associated with a 0.03 decrease in relative risk of spine fractures (95% CI, 0.02–0.05, P= .002). The total treatment effect was a 45% reduction in vertebral fracture risk, but the reduction in BMD was only expected to reduce fracture risks by 20% (relative risk [RR] = 0.8). A separate analysis of the Fracture Intervention Trial (FIT) of alendronate treatment showed that the 3.9% increase in BMD after 1 year of alendronate therapy explained only 16% of the total decrease in the risk of vertebral fracture.
Similar findings were reported for trials on risedronate. Watts et al. conducted 2 meta-analyses (68;69) using individual patient data from 3 pivotal double-blind, placebo controlled trials on risedronate (2.5 mg or 5 mg) with a follow-up period of up to 3 years. The 3 trials are the Vertebral Efficacy with Risedronate Therapy North America (VERT-NA), the VERT Multinational (VERT-MN), and the Hip Intervention Trial (HIP). The first meta-analysis examined the relationship between changes in BMD and vertebral fracture risk reduction. The analysis showed that patients who had an increase in BMD had lower vertebral fracture risk than patients showing a decrease in BMD. However, the reduction in vertebral fracture risk was similar for patients who had a less than 5% increase in BMD and those who had an increase in BMD of 5% or greater. Watts et al. (68) reported that changes in lumbar spine BMD only explained 18% and changes in hip BMD explained 11% of the raloxifene treatment effect. The second meta-analysis explored the relationship between BMD changes and reduction in nonvertebral fracture risk. The analysis showed a similar incidence of nonvertebral fractures for people who had an increase in spine BMD (6.4%) and people who had a decrease in spine BMD (7.8%). Incidence of nonvertebral fractures was also similar regardless of an increase or decrease in femoral neck BMD (7.6% vs 7.5%). Watts et al. (69) determined that changes in spine BMD explained 12% and changes in femoral neck BMD explained 7% of the raloxifene treatment effect on nonvertebral fracture risk.
Two reports published after the above meta-analyses also suggest that changes in BMD do not predict changes in fracture risks (Appendix 13). The Multiple Outcome Raloxifene Evaluation (MORE Trial) (70) enrolled 7,705 postmenopausal women with osteoporosis randomized either to raloxifene treatment (60mg or 120 mg/day) or to a placebo, for 3 years. DXA BMD of the lumbar spine and femoral neck was measured at baseline and annually during follow-up. Sarkar et al. (70) reported that at 3 years, the incidence of fractures was the same regardless of the change in follow-up BMD (0.3%, 3.13% or 6.06%). Even patients with no change in BMD showed a reduction in fracture risk. The percentage change in BMD only accounted for 4% of the observed vertebral risk reduction, while the other 96% of risk reduction remained unexplained. (70)
Chapurlat et al. (71) conducted a post hoc analysis of the relationship between BMD changes and risk of vertebral fracture in women who adhered to alendronate therapy or a placebo. Changes in vertebral fractures were compared among patients who had gained BMD (0% – 4%, or >4%) and those who had lost BMD (lost 0%–4%, or lost >4%) at 1 year. The analyses showed that after one year of alendronate therapy, women who had lost 0 to 4% of BMD at the hip or the lumbar spine compared with the controls had substantial reduction in the risk of vertebral fractures (OR 0.47 & 0.40 respectively), similar to women who had gained 0 to 4% BMD (OR 0.49 & 0.49 respectively) during alendronate therapy. Women who had lost BMD at both the hip and lumbar spine (OR, 0.71; 95% CI, 0.26–1.93) and those who had lost more than 4% BMD during alendronate therapy did not appear to have statistically significant reduction in the risk of vertebral fracture. (71) These findings were not consistent with an earlier analysis of the same trial by Hochberg et al. (72) showing that patients who had gained at least 3% in spine BMD over 12 months of alendronate therapy had the lowest incidence of new vertebral fractures at 36 to 48 months, while those whose BMD remained stable or declined had the greatest fracture risk. (72)
Among clinical trials, the percentage of patients with no significant increase in BMD during drug therapy ranged from 4% to 14%. It was reported that about 10% of elderly patients treated with bisphosphonates lost BMD more than the LSC (Table 17). (73)
Consistent compliance with osteoporosis medication is required to reduce the risk of fractures, but compliance with such medications is low. (74;75) Four observational studies examined the role of BMD testing in influencing patient compliance with osteoporotic drug therapy. These are summarized in Table 18.
Solomon et al. (76) found that a BMD test before or after initiation of drug therapy, along with younger age, less comorbidity, and a fracture, were factors that increased compliance with osteoporosis drug therapy. However, other unknown factors appear to influence compliance. Pickney et al. (77) reported that patients who can recall their BMD results accurately were more likely to have been prescribed an osteoporosis medication and to have remained on the initial medication. Rossini et al. (78) reported that more than 50% of patients who discontinued their osteoporosis medication did so in the first 6-months. They also reported that a BMD T-score of –2.5, a previous vertebral fracture, and corticosteroid treatment were associated with persistence with osteoporosis drug treatment.
Clowes et al. (79) compared compliance to raloxifene in women randomized to no monitoring (n=25), nurse monitoring (n=25) or marker monitoring (n=25). Compliance and persistence with drug therapy was monitored electronically without the patients’ knowledge using a device on the container of the drug. The results showed that patients who were followed up by the nurse had significantly better adherence at one year compared with those without monitoring (increased cumulative adherence to therapy by 27%, P=0.04). Monitoring by marker did not result in an additional improvement in adherence or persistence to therapy compared with nurse-monitoring alone. Patients who were given information of a good response to treatment (measured using change in hip and spine BMD and bone turnover marker) showed significantly better compliance with treatment. The highest frequency of nonadherence occurred in the first 3 months of therapy.
The frequency of BMD monitoring during treatment would depend on the expected rate of increase in BMD in response to treatment and the precision of the test. If the change in BMD is not expected to exceed the LSC in the first year, repeating the test after one year of treatment probably is not likely to provide meaningful results.
A systematic review was conducted on the rate of change in BMD during drug therapy for osteoporosis. Twenty-nine RCTs on etidronate, alendronate, risedronate, parathyroid hormone and HRT that provided data on BMD changes from baseline in the treatment arm were included (Appendices 14 & 15). Mean percent BMD changes from baseline are summarized in Appendix 16. Some of these results are also presented graphically in Figures 13 to to16.16. The largest increase in BMD with treatment occurred during the first year of treatment. However, the increase in BMD even during the first year rarely exceeded 5.54%, the LSC for a BMD test with a 2% precision.
In the RCT comparing patients receiving 400 mg etidronate daily for 14 days followed by 76 days of 500 mg elemental calcium daily to patients receiving 500 mg elemental calcium alone, the etidronate group showed an increase in BMD from baseline of 5.67% in the spine and 1.44% in the femoral neck at 3-years follow-up. The BMD increase from baseline at 1 year was 3.7% at the spine and 1.02% at the femoral neck. Although these changes exceeded the LSC in the research (0.78% based on the reported CV of 0.28%), (80) it is doubtful whether this value will exceed the LSCs of BMD tests in the clinical setting.
Figures 15A and 15D illustrate changes in lumbar spine and hip BMD during 10 years of treatment with alendronate. Only the mean increase in BMD for patients on 20 mg of alendronate exceeded the LSC after the first year. The mean BMD in patients treated with 10 mg of alendronate exceeded the LSC only after 2 years of treatment, and patients receiving 5 mg alendronate after 3 years of treatment. (82)
Among raloxifene, risedronate, parathyroid hormone or strontium ranelate, only strontium ranelate resulted in an increase in BMD close to LSC of a 2% precision test after the first year of treatment (Figure 16). (83)
Although most of the drug studies showed that it took more than 1 year to achieve an increase in BMD greater than a LSC of 5.4%, there were a few exceptions. Evio et al. (84) reported that treatment of postmenopausal women with 10 mg alendronate resulted in a 6.8% increase in spine BMD from baseline after 1 year of therapy, and a combination of alendronate and HRT resulted in an increase of 8.4% from baseline BMD. Black et al. (85) reported a 6.3% increase in the lumbar spine BMD in postmenopausal osteoporotic women after 1 year of treatment with parathyroid hormone, and a 6.1% increase after 1 year of treatment with combined parathyroid and alendronate. Ringe et al. (86) reported an increase of 8% from baseline lumbar spine BMD in osteoporotic men after 1 year of treatment with 10 mg of alendronate.
Cummings et al. (55) reported on the findings of BMD monitoring in 2 RCTs, the MORE Trial and the FIT trial. The FIT trial is a multicenter RCT that compared postmenopausal women with femoral neck BMD of 0.68 g/cm2 or less assigned to treatment of alendronate sodium, and a similar cohort being treated with a placebo. The MORE Trial included postmenopausal women aged 80 years or younger with femoral neck BMD T-score of-2.5 or less or who had at least 1 moderate or 2 mild vertebral fractures detected by spine radiographs. Study subjects were randomly assigned to treatment with 60 or 120 mg/day of raloxifene hydrochloride or an identical looking placebo. BMD was monitored annually for 2 years in the lumbar spine and hip in the FIT study and in the femoral neck in the MORE study. Cummings et al. (55) analyzed changes in BMD in both studies with a total of 3,954 patients. The results showed that in both studies, women with the greatest loss of BMD during the first year of treatment were the most likely to gain BMD during the second year of treatment. Women taking alendronate whose hip BMD decreased by more than 4% during the first year, 83% (95% CI, 82% – 84%) had increases in hip BMD during the second year, with an overall mean increase of 4.7%. In contrast, those women who seemed to gain at least 8% during the first year lost an average of 1% (85% CI, 0.1%–1.9%) during the second year. Similar results were observed among women taking raloxifene for 2 years. Cummings et al. (55) attributed this phenomenon to the principle of regression to the mean, a natural correction of random error in the earlier estimation of change in BMD. In regression to the mean, individuals who have measurements that differ from the mean of a population tend to have repeat measurements that are closer to the mean, and this tendency is greatest for measurements that are farther from the mean.
Experts consulted by the Medical Advisory Secretariat indicated that even though BMD may not be a perfect surrogate for reduction in fracture risk and monitoring response to osteoporosis therapy, BMD testing is presently the most reliable test available for this purpose. The test serves a few purposes. It enables practitioners to identify people who are losing BMD during treatment and to assess whether the patient is taking the medication appropriately. The test also enables physicians to make adjustments to the treatment regimen as required. Bone-sparing drugs such as bisphosphonates are difficult to take and feedback on BMD changes plays an important role in motivating patients to continue with osteoporosis treatment.
Analysis of BMD utilization data in Ontario showed that less than 20% of people 65 years of age and older who had a fragility fracture underwent BMD testing, and only 40% received osteoporosis treatment. About half of the people who received treatment did not have a BMD test. The percentage of men undergoing BMD assessment after a fracture was even lower (about 10%). These findings are consistent with results of studies performed in Canada as well as in other jurisdictions. (Appendix 17)
A systematic review by Papaioannou et al. (87) reported that the prevalence of BMD testing after a fragility fracture ranged from 22% in an Ontario community fracture clinic at 1 year after fracture, to 26% after rehabilitation in an Edmonton study. A retrospective population-based study conducted in the province of Quebec showed that the rate of BMD testing after a fragility fracture was approximately 5% in men and 13% in women, while the rate of treatment was 10% in men and 30% in women. (88)
Low rates of BMD testing and/or treatment after fractures were reported for many countries including the United States, (89) the United Kingdom, (90) and France. (91) A systematic review conducted by Elliot-Gibson et al. (51) that included 37 observational studies, reported that the rate of investigation after a fragility fracture, primarily using BMD measurements, ranged from 0.5 to 32% (median 11%). Only 25% of studies reported treatment rates greater than 10%. Giangregorio et al. (92) conducted a systematic review of 35 studies and reported that BMD was measured in 1 to 32% of patients, and laboratory tests were performed in 1 to 49% of patients for investigation of osteoporosis following a fracture. A diagnosi of osteoporosis diagnosis made in 1 to 45% of patients. A subsequent refracture occurred in 1 to 22% of patients during 6-month to 5-year follow-up. (92)
In order to determine the importance of BMD testing and osteoporosis treatment after a fragility fracture, the Medical Advisory Secretariat conducted 3 systematic reviews to examine (1) the impact of a fragility fracture on the risk of future fractures, (2) the impact of a BMD on the likelihood of osteoporosis treatment and (3) the effectiveness of treatment in reducing the risk of fractures.
Accelerated bone mineral loss of 5.4% from the contralateral femoral neck and 2.4% from the lumbar spine in the year following a fracture has been reported. (93) Since fractures are the most important consequence of low BMD, a systematic review was conducted to examine the relationship between a previous fracture and risk of subsequent fractures. The review included 3 meta-analysis, (94-96) and 13 observational studies published after the meta-analyses. The meta-analyses and current review included large international multicenter population studies with one study exceeding 200,000 people. The studies are mainly prospective longitudinal studies with follow-up periods ranging from 5 to 10 years. The description of these studies is summarized in Appendix 18, and the findings of the studies regarding the impact of a previous fracture on risks of subsequent fracture risks is summarized separately for women and for men in Tables 19 and and2020.
Klotzbuecher et al. (94), conducted a literature review and pooled statistical analysis of the risk of future fracture in people who had a history of prior fracture. The analysis included 33 studies, including prospective cohort studies, case-controlled studies, and cross-sectional surveys published between January 1996 and September 1999. The analysis showed that in women, a prevalent wrist fracture increased the risk of future wrist fractures 3-fold (RR, 3.3; 95% CI, 2.0–5.3), risk of hip fracture 2-fold (RR 1.9, 95% CI 1.6–2.2]), and vertebral fracture by 70% (RR, 1.7; 95% CI, 1.4–2.1). A prevalent vertebral fracture was the strongest predictor for future vertebral fractures, increasing the risk by more than 4-fold (RR, 4.4;, 95% CI, 3.6–5.4). A prevalent vertebral fracture also increased the risk for future hip fracture (RR, 2.3; 95% CI, 2.0–2.87) and wrist fracture (RR, 1.4; [95% CI, 1.2–1.7]). A prevalent hip fracture increased the risk of all fractures by 2-fold or more (RR 2.5 for vertebral fracture, 2.3 for hip fracture and 2.4 for pooled fractures). Klozbuecher reported that in men, a wrist fracture increased the risk of vertebral fractures (RR, 3.3–10) and all incident fractures (RR, 1.8–2.5).
Similar results were reported by Kanis et al., (95) who conducted a pooled analysis of 11 large international prospective cohort studies including 877 men and 4,686 women with 250,000 person years. A previous fracture was associated with a significant increased risk of any subsequent fracture, osteoporotic fracture, and hip fracture at all ages compared with people without a prior fracture. Men and women had similar risk, with relative ratios ranging from 1.93 to 2.30 for men, and from 1.77 to 1.85 for women. The risk ratio was stable with age except in the case of hip fracture where the RR decreased significantly with age.
Haentjens et al. (96) conducted a pooled analysis of 9 cohort studies (1982–2001) to determine the relative risks of subsequent hip fracture in elderly men (>/=50 years) and postmenopausal women who had suffered a Colle’s fracture or spine fracture. The analysis included studies with sample sizes between 36 to 1,905 and follow-up periods ranging from 241 to 40,832 person years. The results were consistent with those of both Klozbuecher et al. and Kanis et al. The impact of a spine fracture on future hip fractures did not differ significantly between genders (RR 3.54 for men vs. 2.20 for women, P= .11). Fractures of the distal part of the radius increased the relative risk of hip fractures significantly more in men than in women (RR 3.26 in men vs. 1.53 in women, PP = .002).
Results of primary studies published since the above systematic reviews lent support to the above findings.
Papaioannou et al. (19) analyzed data for 5,143 postmenopausal women from the Canadian Multicentre Osteoporosis study and reported that a personal history of prior fracture was one of seven independent predictors for incident vertebral and nonvertebral fractures. Prior vertebral fracture appeared to increase the likelihood of developing a clinical vertebral fracture, but was not associated with nonvertebral fractures. A prior forearm fracture appeared to have a greater association with developing an incident main nonvertebral fracture (RR, 3.626; 95% CI, 1.876 –7.008) as compare with any other incident nonvertebral fracture, and is a superior predictor of all incident nonvertebral fractures (RR, 2.521; 95% CI, 1.442– 4.409).
Similar results were reported by Szulc et al. (97) for the Osteoporosis Study in Men sponsored by the French National Institute of Health and Research (MINOS) which included 759 men 50 years of age or older, followed for a mean of 7.5 years. In this cohort, prevalent fractures were associated with a 2-fold increase in the risk of incident fracture (OR 1.28–1.89) when adjusted for age, weight, and BMD, regardless of the site of measurement. For example, for total hip, the OR was 2.07 (95% CI, 1.15–3.76, P = 2.07).
Johnell et al. (98) reported that fracture risk was significantly higher than the general population immediately after a spine, hip or shoulder fracture, especially in younger men (60 years) in whom the RR for new hip fracture reached 125 and RR for or new forearm fracture was 43 immediately after a shoulder fracture. Johnell et al. found that the risk decreased with time.
In addition to fractures, vertebral deformity (usually defined as a reduction of at least 3 SDs in vertebral height from the same-sex normal) has also been found to increase the risk of fractures. Vertebral deformity is a hallmark of osteoporosis and affects at least 20% of the elderly population. (99)
The multicenter European Vertebral Osteoporosis Study (EVOS) (100), a study included in the meta-analysis by Kanis et al., followed 6,344 men and 6,788 women aged 50 years and older from 31 European centers for a median follow-up of 3 years. In this study, Ismail et al. (100) reported that baseline prevalent vertebral deformity was associated with a significant increase in the risk of subsequent hip fracture in women (RR, 4.5; 95% CI, 2.1–9.4), but not in men. The study also found that increasing number of vertebral deformities was associated with an increased risk of all types of limb fractures except distal forearm fracture in both men and women. (Ismail 2001)
A smaller Swedish cohort (101) from the EVOS study (described above), consisting of 298 men and 300 women followed for 10 years, reported that a prevalent vertebral deformity significantly predicted future fractures of any type in both men (age-adjusted HR, 2.7; 95% CI, 1.4–5.3), and in women (age-adjusted HR, 1.8; 95% CI, 1.1–2.9) compared with no vertebral deformity. The predictive value of a prevalent vertebral deformity remained significant after adjusting for age, weight, alcohol consumption, smoking, general health, and previous hip fracture. (101)
Pongchaiyakul et al. (99) conducted a prospective longitudinal study of 114 men and 186 women from the Dubbo Osteoporosis Epidemiology Study (DOES). During 10-year follow-up, people with baseline vertebral deformity had a significantly higher incidence of subsequent fractures (44%) than those without baseline vertebral deformity (27%) with an adjusted relative hazard for any fracture of 2.5 (95% CI, 1.5– 3.9). However, after adjusting for age, sex, and body weight, the effect is only statistically significant for any fractures in women (HR 3.1, 95% CI 1.8–5.4) and in vertebral fractures in both sexes, with an adjusted HR of 5.5 (95% CI, 1.3–22.4) in men and an adjusted HR of 11.1 (95% CI, 3.8–32.3) in women.
A fragility fracture at the hip, spine, or wrist is an independent predictor of subsequent fractures not only at the site of the prevalent fracture, but also at other skeletal sites. The presence of a prevalent fracture increases the risk for any incident fractures by approximately 2 fold or more.
Some guidelines have indicated that BMD measurements should only be made if the results will be used to make treatment decisions. Hence the availability of effective treatment for osteoporosis is an important factor for BMD assessments. The ultimate goal of treating osteoporosis is to reduce the risk of future fractures. The Medical Advisory Secretariat reviewed the evidence relating to the effectiveness of osteoporosis drugs listed in the ODB program. These are HRT, antiresorptive bisphosphonates (e.g. etidronate, alendronate, and risedronate), estrogen receptor modifier (e.g. raloxifene), calcitonin, calcium supplements, vitamin D supplements, and parathyroid hormone therapy.
The Osteoporosis Methodology Group and the Osteoporosis Research Advisory Group conducted 9 systematic reviews and meta-analyses of RCTs (107-116) on osteoporosis therapies for postmenopausal women, comparing osteoporotic drugs with placebo or calcium/vitamin D supplements. There were no direct comparisons between treatments. Quality assessment of the RCTs is summarized in Table 21.
Results of the meta-analyses are summarized in Table 22. Tests for heterogeneity were not statistically significant for studies in alendronate, risedronate, and etidronate, indicating consistent results from study to study. The vitamin D results are relatively consistent. Due to potential publication bias for studies in calcitonin with the pooled estimate driven by 3 small trials with large RR reductions, the RR reduction from a large RCT was presented instead of the pooled estimate. (116)
There was statistically significant reduction in the pooled RR for vertebral fractures with Vitamin D, alendronate, etidronate, risedronate, raloxifene and calcitonin compared with placebo, with RRs ranging from 0.6 to 0.79 (Table 24). (107-110;112;114) The CIs around the pooled estimates suggest that the relative risk reduction is unlikely to be less than one-third for alendronate and unlikely to be less than a quarter for risedronate and raloxifene. Calcium supplements and HRT show trends toward reduction in vertebral fracture, but did not reach statistical significance. (117) (116) The number needed to treat (NNT) to prevent one vertebral fracture in the high-risk population was 72 for alendronate and 94–99 for vitamin D, etidronate, risedronate, and raloxifene. (116)
Alendronate and risedronate were the only 2 drugs that had a significant pooled treatment effect on nonvertebral fracture reduction with RRs of 0.51 and 0.73 respectively (P< .01). The treatment effects were very similar with alendronate across all fracture types. To prevent one nonvertebral fracture in the high-risk population, the NNT was 24 for alendronate, and 43 for risedronate. (116)
With the exceptions of vitamin D and calcitonin, all drugs showed a significant increase in lumbar and spine BMD compared with placebo. The largest treatment effects on the lumbar spine BMD were seen with alendronate and HRT, with intermediate effects seen with risedronate and etidronate. Alendronate, raloxifene, calcium, risedronate, and HRT showed convincing, relatively large effects on BMD in all sites compared with placebo. Greater increases in BMD were observed with higher doses of risedronate, alendronate, and HRT. A dose effect was not observed for calcium or calcitonin. (116)
There were fewer studies on the efficacy of osteoporosis treatment in men that used fracture risk reduction as an outcome. One meta-analysis on alendronate and 2 RCTs on risedronate are summarized in Table 23.
Sawka et al. (118) conducted a meta-analysis on 2 RCTs that compared the effect of alendronate to alfacalciferol or calcium supplements as a treatment of primary osteoporosis in men. One of the studies was a double- blind study (sample size = 77) (119) and the other was an open-label RCT with a sample size of 118. (120) The analysis showed that alendronate significantly reduced the incidence of vertebral fracture in men with primary osteoporosis, with an OR of 0.44 (95% CI, 0.23–0.83). The OR for nonvertebral fracture was 0.6, but this did not reach statistical significance (95% CI, 0.29–1.44). (Table 23)
In an RCT on risedronate, Sato et al. (121) reported a significant risk reduction of hip fracture (RR 0.19, [95% CI,0.04–0.89) with a mean NNT of 16 (95% CI, 9–32). Ringe et al. (86) reported in a small RCT that risedronate significantly reduced the incidence of vertebral fracture from 12.7 to 5.1% (P = 0.28) while there was no significant reduction in the risk of nonvertebral fractures. (Table 23)
No RCTs were found regarding the effectiveness of etidronate or raloxifene in osteoporotic men.
Upper gastrointestinal (GI) events:
The most common adverse event is upper GI side effects. The pooled RR for GI effects in clinical trials of bisphosphonates compared with controls was not statistically significant (e.g. for alendronate, RR = 1.03, 95% CI, 0.98–1.07, P = 0.23). The pooled RR of discontinuing medication as a result of adverse effects from alendronate in 9 clinical trials was also not statistically significant.(108) An endoscopy study (123) on patients receiving risedronate or alendronate reported that overall, gastric ulcers 3 mm and over were observed in 6% of 300 patients on risedronate and 12% of 297 patients during treatment with alendronate. Helitobacter pylori infection did not increase the incidence of bisphosphonate-related gastric ulcers. In the same study, upper GI events were reported by 5.7% of subjects in the risedronate group, and 8.8% in the alendronate group. The symptoms did not predict the presence of mucosal damage.
Osteonecrosis of the jaws
Since 2004, osteonecrosis of the jaws relating to bisphosphonates therapy has been reported. Woo et al. (124) conducted a systematic review of 10 case series on this complication in 2006. A total of 368 cases were reported, of which 4.1% were patients receiving bisphosphonates for treatment of osteoporosis, while 95% of the cases had multiple myeloma and metastatic carcinoma. These cases manifested as exposure of portions of the bone of the mandible alone (65%), maxilla only (26%), or both. Most lesions were on the posterior lingual mandible near the mylophoid ridge and 60% of the cases occurred after a tooth extraction or other dentoalveolar surgery. Most of the published cases (94%) involved new generation intravenous bisphosphonates such as zoledronic acid and pamidronate, which are not listed on the ODB formulary. For bisphosphonates currently covered by the ODB, oral alendronate accounted for 4.2% and risedronate for 0.3% of the cases. Conservative debridement of necrotic bone, pain control, infection management, use of antimicrobial oral rinses, and withdrawal of bisphosphonates were recommended. (124)
There is high-quality evidence that in postmenopausal osteoporotic women, treatment with alendronate, etidronate, risedronate, raloxifene, calcitonin, and vitamin D significantly reduces the risk of vertebral fractures, and only alendronate and risedronate also reduce the risk of nonvertebral fractures. There is some evidence that alendronate and risedronate are also effective in reducing the risk of vertebral fractures in osteoporotic men. The most common adverse event is upper GI side effects including ulcers. Osteonecrosis of the jaw has been reported with the majority of cases being patients receiving high dose intravenous bisphosphonate therapy in association with cancer therapy. The degree of risk for osteonecrosis in patients taking oral bisphosphonates such as alendronate, for osteoporosis is uncertain and warrants careful monitoring (124)
Since evidence shows that there are treatments that effectively increase BMD and reduce risk of fracture, the next question is whether BMD testing is likely to increase the likelihood of treating people with low BMD and at risk of fractures.
There is evidence from 4 prospective observational studies and 1 retrospective study that people with osteopenia or osteoporosis diagnosed with BMD were more likely to receive counselling for osteoporosis prevention, and treatment for low BMD.
Hamel et al.(125) conducted a Canadian prospective cohort study to determine the impact of low BMD and history of stroke on treatment patterns in 1,300 women (aged >20 years) undergoing their first BMD testing. A questionnaire was administered at baseline and again 3 months after BMD testing. Logistic regression showed that treatment decisions were influenced by BMD testing, but not by a history of fracture. There was a substantial care gap in the treatment of patients with osteoporosis either with bisphosphonate or estrogen. (125)
Fitt et al. (126) conducted a prospective study of 385 women age 50 years or older who were referred to a tertiary care hospital to undergo bone density measurement. The proportion of women with osteoporosis receiving HRT or bisphosphonate therapy increased from 15.2% to 63.3% after diagnosis with densitometry. Independent factors associated with the initiation of either therapy were actual BMD results showing osteoporosis (OR, 7.2; 95% CI, 1.7–30.3), subjects’ perception that their scan showed osteopenia or osteoporosis (OR, 13.5; 95% CI, 4.0–45.5), or they were unclear about the results (OR, 3.4; 95% CI, 1.6-18.8), compared with the perception that the results were normal.
Gallagher et al. (127) conducted a questionnaire survey of 1,004 women (age 40–69 years) regarding receipt of osteoporosis prevention counselling, BMD testing, and information on treatment options. Most of the women with osteopenia or osteoporosis reported receiving information about various treatment options (estrogen therapy, calcium, weight bearing exercise), but only 33% reported communication about pharmaceutical alternatives to estrogen replacement therapy, and 20% about vitamin D supplementation. Multivariate analyses showed that women with multiple risk factors for osteoporosis were not being identified for preventive counselling interventions or BMD testing, and the main trigger to physician counselling of women about osteoporosis and its prevention was an osteopenia or osteoporosis diagnosis (OR range 2.15–5.04). (127)
Pressman et al. (128) analyzed information from the Kaiser Foundation Health Plan Persons database to determine the impact of BMD results on the use of osteoporotic drugs in 8,020 women aged 45 years or older who had BMD testing. Logistic regression was used to explore the association between BMD diagnosis and initiation of drug therapy for osteoporosis including HRT, alendronate, etidronate, raloxifene, and calcitonin, within 6 months after BMD testing. The regression analyses showed that diagnosis of osteopenia or osteoporosis by BMD testing increased the likelihood of initiating osteoporotic treatment by 4-fold and 15-fold respectively. Other factors such as age, high exposure to corticosteroid, and history of osteoporotic fracture also increased the likelihood of treatment initiation, but the association was much weaker. (128)
In a United Kingdom survey (90) of 218 people approximately 3 months after a minimal or medium trauma fracture, factors associated with osteoporosis treatment were explored. Multivariate analysis showed that prior bone density scan was 1 of 2 independent predictive factors for receipt of osteoporosis therapy (OR, 8.9; 95% CI, 3.4–23.3). The other predictive factor was age greater than 50 years (OR, 15.2; 95% CI, 1.9–118).
Evidence from observational studies in Canada and other jurisdictions suggests that patients who have undergone BMD measurements and particularly if a diagnosis of osteoporosis is made, are more likely to be given pharmacologic bone-sparing therapy.
Based on a survey of practitioners, Elliot-Gibson et al. (51) reported that barriers to post fracture osteoporosis investigation and treatment cited by physicians in Canada and Ireland were cost of therapy, patient reluctance, time and cost of diagnosing osteoporosis, side effects of medication, lack of access to BMD testing, and lack of time to address secondary prevention. Moreover, until recently, many orthopaedic surgeons did not feel that osteoporosis was their responsibility and, therefore, did not investigate or treat this disease.
Ridout and Hawker (129) conducted a survey of 457 family physicians in province of Ontario (Canada) using a self-administered questionnaire, to examine the use of bone densitometry in the primary care setting. The results showed that few Ontario physicians were significantly limited in their use of BMD. The most often cited reasons for ordering the test were presence of risk factors for osteoporosis (79.4%) and decision-making for HRT (Table 24). The only significant limitations to use identified by more than 10% of respondents were travel distance to a densitometer, and concerns regarding the cost of the test. Results also suggest a positive correlation between the frequency of BMD and the physicians’ reported confidence in the use of the test (correlation coefficient r = 0.25, P = .001). Components of a BMD report perceived to be most useful by physicians were the statement of fracture risk, comparison with age-matched controls and suggestion for investigation and management, supporting the inclusion of clinical data in BMD reports.
Jaglal et al. (130) conducted a mailed survey of a stratified random sample of 1,000 Ontario family physicians from the College of Family Physicians’ database. Three hundred and sixty-four practicing respondents (182 male, 182 females) completed the full questionnaire. There were no statistically significant differences in responses by gender or region of practice. More than 80% of family physicians wanted to be more informed about BMD testing and the pharmacological and nonpharmacological management of osteoporosis. The presence of risk factors was one of the most influential reasons (72%) for ordering BMD testing. (Table 24) Information in peer-reviewed journals was thought to be the most credible. More than 80% were interested in a decision aid that incorporates information on risk factors, fracture risk and a treatment algorithm.
Solomon et al. (131) conducted a cross-sectional survey of 494 physicians in 6 New England states to identify factors associated with ordering few BMD scans. The cohort included physicians in general/family practice, internal medicine, and obstetrics/gynaecology. The mean number of self-reported BMD referrals was 10 (SD11) (median 7) per month. In adjusted logistic models, several factors were found to be significantly associated with referring fewer than 4 patients per month for BMD scans. Internists and family practice physicians, physicians who practised in an urban or rural/small town setting, physicians who spent less than 50% of their time in patient care, and physicians who saw a low proportion of postmenopausal women, were more likely to report ordering fewer BMD tests. Physicians who believed that calcium and vitamin D alone are adequate treatment for osteoporosis and that osteoporosis treatment should not be based on BMD measurements also reported ordering fewer BMD tests. Solomon et al. suggests that the above factors should help provide a rational basis for designing educational strategies aimed at physicians.
Other studies suggest that the beliefs of the orthopaedic surgeon are important determinants of BMD testing and osteoporosis treatment after fractures.
Khandwala et al. (132) surveyed 5 orthopaedic surgeons in Saskatchewan who indicated that osteoporosis treatment was not initiated mainly because they would not be involved in the postoperative follow-up of these patients, and that they believed medical treatment of osteoporosis was the sole responsibility of the primary care physicians.
According to a survey (133) of 3,422 orthopaedic surgeons in France, Germany, Italy, Spain, the United Kingdom, and New Zealand, less than one-fifth of the orthopaedic surgeons arranged for a surgically treated patient with a fragility fracture to have a BMD test. Twenty per cent said that they never refer a patient after a fragility fracture for BMD tests. Only half of the orthopaedic surgeons in Southern Europe know about the importance of some external risk factors for hip fractures. (133)
In a questionnaire survey (4) of 117 orthopaedic surgeons and 113 family physicians in the United Kingdom, 81% of the orthopaedic surgeons and 96% of family physicians agreed that low trauma fractures in patients over 50 years old required investigation for osteoporosis. However, only a small percentage of orthopaedic surgeons would routinely assess and start treatment for osteoporosis or refer to an osteoporosis clinic in patients over 50 years old following a Colle’s fracture or a femoral neck fracture (17%). Similarly, without prompting from the orthopaedic surgeon, only 33% of general practitioners would routinely investigate for osteoporosis after a Colle’s fracture and 38% after a femoral neck fracture. Prompting from the orthopaedic surgeons would increase osteoporosis investigation by 22% for Colle’s fracture and 21% for femoral neck fracture. In comparison to wrist fracture and femoral neck fracture, patients with vertebral wedge fractures are relatively well investigated and treated by orthopaedic surgeons (71%) and general practitioners (64%). (4)
Surveys of physicians and patients suggest that the following are factors that may increase the appropriate use of BMD in high-risk patients:
Patients with fragility fracture represent a target population in whom the use of BMD testing and subsequent treatment can be optimized. Since orthopaedic surgeons are usually the first medical practitioner to see osteoporotic fractures, they play a crucial role in initiating osteoporosis investigation and treatment in patients with a fragility fracture. The World Orthopaedic Osteoporosis Organization strongly advocates a leading role for orthopaedic surgeons in the management of osteoporosis in their fragility fracture patients. (51;134)
Many initiatives have been developed in Canada and other jurisdictions to improve the investigation and treatment of osteoporosis in patients after a fragility fracture. A few examples are provided below.
Factors that predispose men to fragility fractures are less well established than in women. The follow sections reviewed 7 systematic reviews and 24 observational studies that pertain to predictors of fragility fractures in men. These studies are summarized in Appendices 19 to 20.
The clinical use of bone densitometry is in the relation between BMD and fracture risk. BMD has been found to be one of the most important determinants of fractures in women. The effectiveness of a BMD test depends on its ability to predict fracture risk, often expressed as the gradient of risk, which is the RR of fracture for every standard deviation decrease in age-adjusted mean BMD (RR/SD). The larger the gradient of risk, the higher is the predictive value. One meta-analysis and 7 studies addressed the predictive value of BMD in men (Table 25).
The meta-analysis by Johnell et al. (142) was based on individual data from 12 cohort studies conducted in Europe, North America, Australia, and Asia, consisting of 38,973 patients (75% female) with a follow-up of 168,366 person-years. The characteristics of these studies are summarized in Table 26.
BMD was assessed at the femoral neck by DXA. The Z-score for each cohort was computed from the regression of BMD by age. The results showed no difference in the gradient of risk (predictive power) afforded by BMD at the femoral neck between men and women. The gradients of risk were highest for the prediction of hip fracture, lowest for any fracture, and intermediate for osteoporotic fracture. For hip fracture risk, the gradient of risk per SD was marginally higher in men than in women, but this was not apparent when gradients of risk were examined by unit of BMD or by age. (Table 27)
The study showed that for any fracture and for osteoporotic fractures, the gradient of risk increased significantly with age in both men and women, reaching a plateau at about age 80 (Figure 17). For the prediction of hip fracture, the gradient of risk decreased with age, with no differences between men and women (Figure 18). However, the absolute risk still rose markedly with age. The age-specific risk of hip fracture at a given hip BMD in men was the same in women with the same BMD and age. (142)
For predicting any fractures or osteoporotic fractures, there was a higher gradient of risk the lower the baseline BMD. For example at a baseline Z-score of –4, the relative risk for osteoporotic fracture was 2.1 per SD (95% CI, 1.63 – 2.71) and at a Z-score of –1, the relative risk was 1.73 per SD (95% CI, 1.59– 1.89). The baseline Z-score did not have a significant impact on the gradient of risk for predicting hip fractures (Figure 18).
Gradients of risk did not change as time elapsed after BMD measurement for all fractures, but had a nonsignificant attenuation for predicting osteoporotic fracture or hip fracture. The decrease in predictive ability was small and did not markedly affect the computation of 10-year fracture probability. (142)
The 7 studies not included in the meta-analysis by Johnell et al. (142) included 3 population-based longitudinal studies, 3 cross-sectional studies, and 1 case-controlled study (Table 25). Three of the studies included only men and the other 5 included both men and women. The number of men included in the studies ranged from 62 (146) to 22,444. (145) Follow-up periods of the longitudinal studies ranged from 6.3 years to 16 years. Most of the studies expressed the relationship between BMD and fracture risk in RR, OR or HR for fracture per SD decrease in T-score or Z-score.
Van der Klift et al.,(144) Gonnelli et al., (143) Schuit et al., (145) and Pande et al., (146) all reported that low BMD increased the risk of fragility fractures (Table 25). The reported mean predictive values of BMD in men, expressed as RR/SD, OR/SD, or HR/SD decrease in BMD ranged from 2.42 to 3.42 for the hip and 1.28 to 2.3 for the spine (Table 25). Cauley et al. (147) reported that a 0.10 g/cm2 decrease in real BMD was associated with a 30 to 40% increased odds of a vertebral fracture in men compared with a 60 to 90% increased odds in women.
In the MINOS study, Szulc et al. (97) provided accuracy data based on BMD cut-offs. This study followed 759 French men aged 50 years or more for 90 months to compare the predictive value of BMD T-scores of –2.0 and –2.5 at different sites for osteoporotic fracture. The study found that BMD was predictive of osteoporotic fractures at all sites with ORs varying from 1.28 to 1.89 per 1 SD decrease in BMD (P< .05– .0001). The sensitivity and specificity of the 2 BMD thresholds are summarized in Table 28.
The above data shows that despite a strong association between BMD and fracture risk, the sensitivity of BMD to detect men at high risk of fracture is low. Only 14 to 45% of fractures were observed in men with a T-score of less than –2, and 27 to 45% of fractures occurred in men with a T-score between –1 and –2. The low sensitivity was explained by the limited number of fractures that occurred in men with low BMD regardless of the measured site. Area under the Receiver Operator Characteristics curve ranged from 0.643 (95% CI, 0.592–0.693) for femoral neck to 0.697 (95% CI, 0.627–0.765) for the distal forearm, indicating that BMD itself has a limited value for detecting individual men who will actually have a fracture in the future. Similarly, Schuit et al. (145) reported that a T-score less than –2.5 identified only 21% of nonvertebral fractures in elderly men, even lower than the 44% in women. (97)
The meta-analysis by Kanis et al. (95) found that both the risk of fracture at a specific age and BMD are similar for both genders, but Cauley et al. (147), based on an analysis of the data from 2 longitudinal studies, found that the areal BMD of men with a vertebral fracture was 20 to 38% greater (P <.05) than the areal BMD of a woman with a fracture. For areal BMD, the curves for men and women had different slopes, suggesting a different probability of fracture at absolute levels of areal BMD. Kudlacek et al. (148) also reported that men fractured at a higher BMD value than women (OR for gender 3.1).
A systematic review is on factors that influences BMD in men is being conducted under a guideline initiative of the Osteoporosis Strategy at the ministry. Hence these factors will not be addressed in this report.
Knowing risk factors other than BMD that predisposes men to fragility fractures will assist in case finding for BMD testing. Hence meta-analysis and studies on predictors (other than BMD) for fractures in men will be reviewed.
The CaMos study (19) (discussed in the Background section) identified low BMD, previous history of fragility fracture, comorbid conditions (kidney disease and inflammatory bowel disease) as significant predictors of fragility fractures in women.
Espallargues et al. (149) conducted a large systematic review to identify factors associated with the development of low bone mass and classify these risk factors according to the strength of their association with fracture incidence. This review included 94 cohort studies, 72 case-controlled studies, and 1 RCT. The quality of the studies was assessed to be moderate. Only 57% of the studies included men compared with 94% for women. Based on both qualitative and quantitative analysis, Espallargues et al. classified risk factors for fractures according to level of risks. Table 29 summarizes the high and moderate-risk factors for fracture-related bone mass loss that have point estimates for RR from the meta-analysis.
Moderate-risk factors for fractures included female sex, active smoking, low sunlight exposure, family history of osteoporotic fracture, and surgical menopause. (149)
This systematic review was not specific to men. Moreover, it could not determine whether these factors were independent predictors of fracture when all risk factors were combined or whether they provide additional information beyond other factors since the results were not based on individual patient data.
Studies relating to risk factors for fragility fractures in men are reviewed in the following sections.
Kanis et al. (23) explored the relationship between 10-year probabilities of fractures, BMD, and age in Swedish men and women. This study showed that age provided an independent element of risk not captured by BMD. In men, forearm risk is stable with age. For other fractures in men, as for all fractures in women, fracture risk increases with age up to 80 years. Thereafter, the 10-year probability plateaued or decreased since mortality exceeded the fracture risk. (23)
The 10-year probability for hip fracture in men with a T-scoreless than or equal to –2.5 at the femoral neck is 5.1% at age 50 and increases to 24.3% at age 80 with the same T-score. At a T score of –2.5, any difference in 10-year fracture probabilities for hip and vertebral fracture between men and women is not marked, since the same BMD measured at the same site at the same age carries a similar fracture risk in both sexes. For example, a femoral neck T-score of –2.5 at age 85 carries a risk of hip fracture of 10.5 for men and 10.0 for women. At the same age, the 10-year probabilities are higher in women than men.. (23)
Szulc et al. (97) reported that in the MINOS study, the incidence of fractures increased with age (OR = 1.29 per 10-year increase, 95% CI, 1.00–1.64, P = .05). In men aged more than 75 years, the fracture incidence was almost 3 times higher than in men aged less than 55 years. The age-related increment of the fracture incidence was mild but significant and independent of BMD. The results also suggest that age itself is not an independent risk factor for fracture, but rather a surrogate for age-related risk factors such as the risk of falls, lower limb disability, or impaired balance.
Cauley e al. (147) also found that the risk of vertebral fracture increased with age in both men and women. The prevalence was 14%, 20%, and 28% among women and 11%, 13%, and 29% among men aged less than 69 yrs, 70-79 years and 80+ years respectively.
Evidence presented in the section on previous fractures showed that a history of previous fractures increases the rate of subsequent fractures in men as well as in women.
Previous studies have shown that low BMI and weight loss are associated with increased bone loss in older men (150) (60) (151). Table 30 summarizes 1 meta-analysis (152) and 7 primary studies that explored the impact of body weight, BMI, and weight loss on the risk of fractures in men.
De Laet et al. (152) conducted a meta-analysis of 12 prospective studies (>6,000 men and women, >250,000 person years) to explore the relationship of BMI with fracture risk in men and women. The effect of BMI, BMD, age and gender on the risk of any fracture, any osteoporotic fracture, and hip fracture were analyzed using a Poisson regression model in each cohort, separately and for merged results. Without information of BMD, low BMI significantly increases the age-adjusted risk of any fractures, osteoporotic fractures, and hip fractures. The effect of BMI on fracture risk is independent of age or gender, but dependent on BMD. The overall risk gradient for men and women was an RR of 0.98 per unit increase in BMI. The RR for fracture risk with BMI was nonlinear, the RR was markedly higher at the lower values of BMI, particularly with a BMI less than or equal to 20 kg/ m2. For example, the risk for hip fractures decreased by 17% (RR 0.83, 95% CI 0.69–0.99) when the BMI increased from 25 kg/m2 to 30 kg/m2. However, the risk for hip fractures doubled (RR 1.95, 95% CI 1.71–2.22) when the BMI decreased from 25 kg/m2 to 20 kg/ m2. De Laet et al. (152) commented that obesity should not be regarded as an important protective factor for hip fracture risk. Rather leanness should be regarded as a significant risk factor for fractures. The authors also suggested that low BMI could be used to enhance the predictive value of BMI in case finding. After adjustment for BMD, BMI was only predictive of hip fracture risk for men and women at a BMI of 20 kg/m2 or less.
The 7 primary studies included 5 longitudinal studies and 2 case-controlled studies. The number of men included in the studies ranged from 192 to 22,444. The follow-up period for the longitudinal studies ranged from 3.8 to 16 years.
A high BMI has been found to protect against hip and pelvis fractures in both men and women. Kelsey et al. (31) reported an OR of 0.65 (95% CI, 0.52–0.81) for pelvis fracture with a 5 unit increase in BMI, while Holmberg et al. (31) reported an age-adjusted relative risk of 0.63 (95% CI, 0.53–0.76, P = .0001) for hip fracture with every SD increase in BMI.
The Mediterranean Osteoporosis Study (MEDOS) (155) compared 730 men aged 50 years and older who had a hip fracture with 1,132 matched controls who did not have a hip fracture. The results showed that a low BMI was associated with a significantly increased risk of hip fracture in a linear dose dependent manner. In univariate analysis, the risk of hip fracture decreased by a mean of 6.8% (95% CI, 4–9) for every unit increase in BMI. The risk decreased significantly with increasing weight but the increase in risk with increasing height was small and not statistically significant. Other protective factors were consumption of cheese, and exposure to sunlight. A high consumption of alcohol and a long duration of smoking increased the risk for fractures. In multivariate analysis, BMI, leisure exercise, exposure to sunlight, and consumption of tea, alcohol, and tobacco remained independent risk factors, accounting for 54% of hip fractures. The use of risk factors to predict hip fractures had relatively low sensitivity (59.6%) and specificity (61.0%). According to the MEDOS study, these potentially modifiable risk factors are similar to those reported in women from the same study. (155)
The European Prospective Osteoporosis Study (EPOS) (153) which followed more than 3,000 men for a mean of 3.8 years, found that only in men was an increase in BMI significantly associated with a reduced risk of incident vertebral fracture as defined qualitatively (RR 0.76 per 1 SD change in BMI adjusted for age and centre, 95% CI, 0.60–0.97). Neither the height nor weight had significantly influenced the risk of vertebral fracture in either gender. Men in the lowest BMI quintile had an increased risk of incident vertebral fracture (RR, 1.99; 95% CI, 1.01–3.93). (153)
Szulc et al. (97) reported in the MINOS study that in men, the unadjusted fracture risk increased with decreasing body weight (OR = 1.15 per 5 kg decrease in body weight, 95% CI, 1.03–1.28, P< .02). The study also found that men in the lowest quartile of body weight (<71 kg), the fracture incidence was twice as high as in men in the highest quartile. Similar to EPOS, body height was not associated with risk of fracture.
Numerous studies have reported an association between weight loss and increased risk of fracture. Two studies that yielded data on men were found. Langlois et al. (157) analyzed data from the 3 sites of the Established Populations for Epidemiologic Studies of the Elderly (EPESE) with 2,413 community-dwelling white men 67 years of age or older, followed for a mean of 8 years. The overall incident rate of hip fractures was 5.3 per 1,000 person-years. Extreme weight loss (>/=10%) beginning at age 50 years in older men was associated with a significant increase in risk of hip fracture. The unadjusted RR of hip fracture among older men with extreme weight loss was 3 times that of both men with lesser weight loss and those with little change in weight. After adjustment for other risk factors for hip fracture including BMI, older men with a weight loss of at least 10% still had a 2-fold increase in risk of fractures (RR, 1.85, 95% CI, 1.04–3.31). Men with a lesser decrease in weight (5% to <10%) did not have a significant increase in risk of fracture. Conversely, men with weight gain of 10% or less had a borderline significant decrease in the risk of hip fracture (RR, 0.38; 95% CI, 0.14–1.00). Men with extreme weight loss were associated with several indicators of poor health, suggesting that weight loss is a marker of frailty that may increase the risk of hip fracture. (157)
Meyer et al. (156) conducted a prospective longitudinal study of 19,151 men and 19,938 women (mean age 49 years) in Norway over a mean follow-up period of 12 years. Weight variability was calculated from 3 consecutive weight measurements during follow-up. The results showed that in both men and women, those people with the most weight variability had an increased risk of hip fracture (RR, 1.24; 95% CI, 1.25–5.86 in men, and RR, 2.07; 95% CI, 1.24–3.46 in women). Overall, the effect of weight variability was not affected by adjustment for BMI and linear trend in weight change. In men, those losing weight also had significantly higher risk of fracture compared wit men gaining weight (RR, 2.01; 95% CI, 1.19 –3.41). (156)
Kanis et al. (158) conducted a meta-analysis of 7 prospective population studies (12,567 men, 22,361 women, and 134,374 person-years) to explore the relationship between family history of fracture and fracture risk. The risk of fracture was estimated by applying Poisson regression to each cohort and each sex separately. Covariates used in the model included time since start of follow-up, age at baseline, family history of fracture, BMD, and the interaction term, current age × family history and BMD. The results of each cohort and the 2 sexes were weighted according to the variance and merged to determine the weighted mean and standard deviation. The meta-analysis showed that in men, a history of any fracture in a parent was associated with a significant increase in risk ratio for hip fracture (RR, 2.02; 95% CI, 1.18– 3.46) and a sibling history was associated with a significant increase in the risk of any fracture (RR, 1.66; 95% CI, 1.23–2.24) and osteoporotic fracture (RR, 1.58; 95% CI, 1.07–2.32). In women, parental history of any fracture was associated with an increased risk of any fractures, osteoporotic fractures and hip fractures, but there was no significant impact from sibling history of fracture. The risk ratios were generally higher in men compared with women, but the difference was not statistically significant. In men, a family history of hip fracture was not associated with a significant increase in risk of fractures whereas in women, parental history of hip fracture was associated with increased risk of any fracture, osteoporotic fracture, and hip fracture. A sibling history of hip fracture was not associated with significant changes in risk of fractures in men or women.
Glucocorticoids decrease intestinal absorption of calcium and phosphate and increase urinary excretion of calcium. In addition, long-term exposure to glucocorticoids inhibits osteoblast proliferation and reduces sex hormone production. The combined result is a loss of BMD reported as high as 8% in the trabecular bone and 2% in the cortical bone of the lumbar spine over a 20-week period at a mean dose of 7.5 mg/day prednisone. (62) No studies that included only men were found. One meta-analysis (159) was found that included 42,542 men and women with 176,000 person-years from 7 prospective population studies (2 including women only). Three case controlled studies, and 5 cohort studies that were not included in the meta-analysis were also reviewed. (Table 31) All included both men and women.
The meta-analysis and primary studies suggest that oral corticosteroid therapy significantly increases the risk of vertebral fractures, nonvertebral fractures, hip fractures, osteoporotic fractures, and all fractures, even after adjustment for BMD. The impact becomes significant at a dose of 7.5 mg for 6 months. (160) The meta-analysis by Kanis et al. suggests greatest impact for hip fractures (RR 2.48-4.42), and in younger men (50 years vs 85 years). A dose response (161;162) and a duration effect (RR 3.27 for vertebral fracture for >90-day therapy vs. RR 2.88 for <90-day therapy) were observed. (162)
Increased risk of any fractures associated with the use of inhaled glucocorticoid therapy was reported (163-165). A slight but significant increase in the risk of hip fractures was also found (RR 1.19–1.26) in people using inhaled glucocorticoid therapy. (166;167) Van Staa et al. reported an increased risk for both vertebral (RR 1.51) and nonvertebral fractures (RR 1.15) in a retrospective cohort study. (166)
Sex hormones are important for the growth and maintenance of the skeleton. In women, reduced serum levels of estradiol are associated with an increased risk of incident fractures. (168;169) However, there was conflicting evidence regarding the relationship between levels of sex hormones and risk of fracture in men.
Barrett-Connor et al. (170) studied 352 men (mean age 66 years) and 288 postmenopausal women (mean age 72 years) in the Rancho Bernardo Study. The results showed that men with at least one vertebral fracture had significantly lower levels of total and bioavailable estradiol with no significant differences for other hormones. There was a graded association between increasing concentrations of total and bioavailable estradiol and decreasing fracture prevalence. Men in the lowest quintile of total or bioavailable estradiol had significantly higher odds for vertebral fracture than those in the highest quintile. The OR for total estradiol was 4.16 (95% CI, 1.22–14.19) and for bioavailable estradiol 5.08 (95% CI, 1.20–21.51). Testosterone levels were not associated with vertebral fractures in men in quintile analysis. In women, vertebral fractures were not associated with any of the hormones or with other covariates including BMI, weight loss, alcohol consumption, current smoking, exercise, current use of thiazide diuretics, thyroid hormones, or calcium supplementation.
Mellstrom et al. (171) explored the relationship of sex hormone levels and self–reported prevalent fractures (after age 50 years) in 2,908 elderly men (mean age 75.4 years) in the cross-sectional Swedish MrOS Study. Mellstrome et al. reported that the free level of testosterone was an independent positive predictor of BMD in total hip, total body, femur trochanter and arm, but not in the lumbar spine. Free estradiol was an independent positive predictor of BMD at all sites, especially the lumbar spine. Free estradiol and free testosterone were stronger predictors of BMD than the respective total sex hormone levels. Free testosterone levels below the median were positive predictors of prevalent fractures after 50 years of age (OR, 1.26; 95% CI, 1.04–1.53, P< .05), osteoporotic fractures (OR, 1.47; 95% CI, 1.09– 1.98, P< .05), and prevalent x-ray confirmed vertebral fractures (OR, 1.85; 95% CI, 1.29–2.66, P< .001). The predictive value of free testosterone was not affected by adjustment for BMD, age, height, weight, smoking status, physical activity, and calcium intake. Free estradiol below the median did not significantly predict any of the fracture-related parameters. However, free estradiol in the lowest 10 percentile was a strong positive predictor of X-ray-verified vertebral fractures (adjusted OR, 2.31; 95% CI, 1.39–3.86, P< .001) and height loss of greater than 5mm (OR, 1.63; 95% CI, 1.14–2.34), suggesting that a threshold level exists for free estradiol to affect bone health. (171)
Kanis et al. (172) conducted a meta-analysis of 3 prospective cohort studies from Canada, Australia, and the Netherlands to quantify, in an international setting, the risk associated with alcohol consumption. The meta-analysis included 16,971 people (5,939 males and 11,032 females with a total follow-up of 75,433 person years). BMD was measured at the femoral neck by DXA at all centres. The risk of fracture was estimated by applying Poisson regression to each cohort and each sex separately. Covariates included current time, current age, alcohol intake, and alcohol intake times current age. There was no significant heterogeneity in risk between cohorts. Intake of alcohol was higher in men than in women (19% of men vs. 4% of women had more than 2 units per day, 8% of men vs. 1% of women took >/= 5 units per day). When assessed as a continuous variable, high intakes of alcohol were associated with an increased risk of osteoporotic fracture or of hip fracture which was not statistically significant (RR for hip fracture 1.07 (95% CI 1–1.3) for men and 1.11 (95% CI 0.98–1.26) for women. When the risk ratio was assessed according to units of alcohol consumed, the risk ratio increased with more than 2 units per day in both men and women. (RR hip fracture = 1.38; 95% CI, 1.21–3.03 for men and RR hip fracture 1.33; 95% CI, 1.01–1.75 for women). The risk ratio increased with higher categories of intake. There was no effect on risk ratio when BMD was added to the model. When intake was dichotomized at more than 2 units daily, there was no confounding effect of smoking or BMI on the association. (A high intake of alcohol confers a significant risk of future fracture, which is over and above that which can be explained by variations in BMD). There was a threshold effect with no increased risk of osteoporosis or hip fracture in individuals who consumed 2 units or less per day of alcohol.
Kanis et al. (173) conducted a meta-analysis of 10 prospective cohort studies to explore the risk associated with smoking on future fractures. The meta-analysis included 59,232 people (43,832 women and 15,400 men). Mean age ranged from 52.3 to 80 years and, based on self-reporting, 18% had a history of current smoking. Current smoking was associated with a significantly increased risk of any kind of fracture including osteoporotic or hip fracture in both men and women. For hip fractures alone, there was no difference in risk ratio between men and women. For men and women combined, risk with current smoking was highest for hip fracture (RR = 1.84), lowest for overall fractures (RR = 1.25) and intermediate for osteoporotic fracture (RR = 1.29). Risk ratio was adjusted downward when taking BMD into account and was no longer significant for osteoporotic fracture in women. In men and women combined, low BMD accounted for 45% of the risk for overall fractures associated with smoking, 40% for osteoporotic fractures, and 23% for hip fractures. In multivariate analysis including BMI, and BMD, the risk ratio for smoking remained significant for overall fractures and for hip fractures. A history of smoking was also associated with a significant risk increase for any fracture (RR, 1.19; 95% CI, 1.07–1.51) and specifically for an osteoporotic (RR, 1.18; 95% CI, 1.09–1.27), or hip fracture (RR, 1.38; 95% CI, 1.15–1.65). In summary, smoking carries a modest but significant risk for future fractures. The effect of smoking is over and above that which can be explained by variations in BMD. The risk was greater for hip fracture than for all fractures and osteoporotic fractures.
The risk factors for fragility fractures in men are similar to those found in women. The risk factors most predictive of fractures in men (relative risk >/=2) are:
Evidence supports the use of BMD in conjunction with assessment of other risks for fractures (age, history of fractures, family history, and body weight, etc.) to determine risk of fractures in men as in women. Efforts are being focused on developing 5 to 10 year probabilities of fractures based on BMD, age and other risk factors.
There are presently no Canadian guidelines for BMD testing that are specific to men. Most guidelines address only women. Osteoporosis Canada’s current guidelines recommend screening for both men and women age 65 years and older as well as for younger individuals (between age 50 and age 64) who have risk factors for fractures. Osteoporosis Canada is preparing for submission to Ontario’s Osteoporosis Strategy new guidelines for BMD testing in men. These have not yet been released. (Personal communication, November 17, 2006)
The 2004 official position of the International Society for Clinical Densitometry (ISCD) (18) recommends BMD testing in all men 70 years of age and older and in younger men who have experienced a fragility fracture or had a condition or are taking medication associated with low BMD. ISCD also provided diagnostic definitions for osteoporosis in men (Table 32).
The section on treatment of osteoporosis showed that, based on evidence from small RCTs, alendronate and risedronate significantly reduce the risk of vertebral fractures in men with osteoporosis, and there is evidence to suggest that risedronate reduces the risk of hip fracture in this population.
In a Canadian burden of illness study on Osteoporosis, Goeree et al. (2) reported that hip fractures caused greater morbidity, higher mortality and more expenditures than any other fractures combined. In addition, approximately one-third of hip fracture patients may become totally dependent on support and in turn become largely dependent on long-term institutionalization for care. They also reported that between 12 to 40% of all hip fracture patients die within 6 months, and the excess mortality rate has been reported to be between 12 to 25% during the first year following a hip fracture.
According to Wiktorowicz et al. (12), the mean one year cost of hip fractures in Canada was estimated at $26,527 (Cdn). However, this cost varied by the patient’s place of residence, age and survival to one year. The authors estimated the average cost (Cdn dollars) of hospitalization for those admitted for a hip fracture to be approximately $21,385 for community residents returned to the community, $44,156 for community residents transferred to long-term care, $33,729 for long-term care residents and $15,498 for those who died within 1 year of hospitalization. They estimated that the annual economic implication of hip fracture in Canada was $650 million in 2001 and was expected to increase to $2.4 billion by 2041. They also reported that among patients who experienced a fracture, the risk of a future fractures increased 20-fold.
In addition to reducing the burdens associated with osteoporosis, one of the objectives of increasing BMD testing is to improve patient management associated with low BMD. A study by Stock et al. (174) suggests that only 30% of physicians who receive a short version of a BMD test report understand it compared with86% of those who receive a more comprehensive BMD test report. The longer reports led to greater modifications in pharmacologic treatment of osteoporosis and less confusion about reports by physicians.
The Medical Advisory Secretariat uses a standardized costing methodology for all of its economic analysis of technologies. The main cost categories and the associated methodology from the province’s perspective are as follows:
Hospital: Ontario Case Costing Initiative (OCCI) cost data is used for all program costs when there are ten or more hospital separations or one-third or more of hospital separations in the ministry’s data warehouse are for the designated ICD-10 diagnosis and CCI procedure codes. Where appropriate, costs are adjusted for both hospital specific or peer-specific effects. In cases where the technology under review falls outside the hospitals that report to the OCCI, PAC-10 weights converted into monetary units are utilized. Adjustments may need to be made to ensure that the relevant Case Mix Group is reflective of the diagnosis and procedures under consideration. Due to the difficulties of estimating indirect costs in hospitals associated with a particular diagnosis/procedure, the MAS normally defaults to considering direct treatment costs only. Historical costs have been adjusted upward by 3% per annum representing a 5% inflation rate assumption less a 2% implicit expectation of efficiency gains by hospitals. Non-Hospital: These include physician services costs obtained from the Provider Services Branch of the Ontario Ministry of Health and Long Term Care, device costs from the perspective of local health care institutions and pharmaceutical costs from the Ontario Drug Benefit formulary list price. Discounting: For all cost-effective analysis, discount rates of 5% and 3% are used as per the Canadian Coordinating Office for Health Technology Assessment (CCOHTA) and the Washington Panel of Cost-Effectiveness, respectively. Downstream cost savings: All cost avoidance and cost savings are based on assumptions of utilization, care patterns, funding and other factors. These may or may not be realized by the system or individual institutions.
In cases where a deviation from this standard is used, an explanation has been given as to the reasons, the assumptions and the revised approach.
The economic analysis represents an estimate only, based on assumptions and costing methodologies that have been explicitly stated above. These estimates will change if different assumptions and costing methodologies are applied for the purpose of developing implementation plans for the technology.
Analyses of Ontario data suggest that only about 19% of people underwent BMD testing and about 40% received treatment after a fragility fracture. The analysis also suggests that people were more likely to be treated with antiresorptive drugs following a fracture if they had undergone BMD testing during the first year after the fracture. There is evidence from RCTs that antiresorptive drugs significantly reduce the risk of fragility fractures. These findings support increasing the use of BMD tests following a fragility fracture. A cost-effectiveness analysis was conducted on increasing the use of BMD testing in and improving reporting for postmenopausal women aged 65 years and older during the first year following a wrist or hip fracture. The following data were used in the analysis.
Based on administrative data, 2,737 males and 10,706 females aged 65 years and older in Ontario experienced either a hip or wrist fracture in 2005. These figures represent approximately 1% of the total population of males and females aged 65 and over in Ontario.
All costs are in Canadian dollars unless specified otherwise.
Total cost of a BMD = $106 (46)
Hospital Costs (Average Costs)
Due to a lack of data from the Ontario Case Costing Initiative, all hospital costs were obtained from a Canadian study on the economics of hip fractures by Wiktorowicz et al. (12) and adjusted to present value using a discounting rate of 5%. Annual hospital costs associated with hip fractures were estimated for 4 separate categories of patients as follows:
The average annual cost of treatment for low BMD with biphosphonates = $391. This estimate was based on weekly doses of 70mgs of alendronate (fosamax) and 35mgs of risedronate sodium (actonel). The cost of a 70mg tablet of alendronate is estimated at $6.195, and the cost of a 35mg tablet of risedronate sodium is estimated at $8.85. (175)
A decision analysis was conducted using TreeAge Pro 2006 software to compare 3 different options for bone mineral densitometry testing among women aged 65 years and older with a previous history of wrist or hip fracture over a one-year time frame. These 6 different options were as follows:
The rationale behind introducing the “improved reporting” arms of the decision analysis is that, according to expert opinion, physicians may lack the ability to interpret the tests and are not able to fully understand them. Improved reporting has the potential to improve management and lead to higher treatment rates in women with low BMD. This would then lead to a reduction in future hip fractures that are the most resource intensive kinds of fractures. According to Cadarette et al., (176) approximately 80% of women in Ontario aged 65 years of age and older with a prior fracture were found to have low BMD upon testing; ideally these women should be on treatment. This objective of improved reporting to capture the targeted 80% of women aged 65 and older with a previous hip or wrist fracture who have low BMD was used to model the “improved reporting” arms of the decision analysis.
The assumptions used in the decision analysis were as follows:
Figure 19 illustrates the different strategies that were considered to determine the most cost-effective strategy for increased BMD testing and reporting.
Table 33 illustrates the probabilities that were used in each of the chance nodes (indicated by green circles) in the decision analysis.
Table 34 summarized the results from the decision analysis for all 6 different strategies under consideration.
A decision analysis showed that increasing the rate of BMD testing following a fragility fracture in people age 65 years and older and improving BMD reports to physicians would be cost-effective, resulting in a cost-effectiveness ratio ranging from $2,285 (Cdn) per fracture avoided (worst-case scenario) to $1,981 (Cdn) per fracture avoided (best-case scenario).
The analysis also showed that approximately $0.3 million to $2.8 million (Cdn) could potentially be saved in annual downstream costs due to a reduction in hip re-fracture rates among women 65 and over with a previous history of hip or wrist fractures, if they were tested for low BMD within one year after a wrist or hip fracture. If the total number of males aged 65 and over were included in the analysis under the same assumptions that were made for females, the total savings would increase to a range of $0.3 million to $3.5 million.
Analysis of Ontario utilization data suggests that baseline and serial BMD tests may have been overused in people at low risk of accelerated bone loss, and underused in high-risk men and in people following a fragility fracture. In order to obtain the full benefit of BMD tests, use of BMD in the province should be shifted from low-risk individuals to those at high risk of osteoporosis and fractures. An analysis was conducted on the budget impact of increasing the rates of BMD testing in high-risk men and in people aged 50 years and over following a hip or wrist fracture, while reducing unnecessary testing and lengthening the interval of serial testing in low-risk individuals. Tables 35 and and3636 summarize these data according to current utilization within each category as well as expected utilization within each category.
At present, there are approximately 35,800 repeat BMDs (within 24-month of the previous test) were performed annually in low risk individuals under the age of 65 years. Analysis in a previous section on repeat BMD in low-risk people not receiving osteoporosis treatment showed that the appropriate interval between two consecutive BMD tests in this population can be between 3 to 5 years, depending on the precision of the test. Increasing the interval of serial testing for low risk individuals from every 2 years (to every 3 to 5 years) could potentially reduce repeat testing by 5,000 to 9,700 per year. For initial and annual testing in low-risk people, the projected utilization was calculated assuming that 50 to 80% of the 2005/06 claims coded as low risk were coded correctly.
The projected increase in BMD tests after a fragility fracture (in people ≥ age 50 years) was estimated based on the assumption of a 50% and 80% increase in BMD testing used in the decision model. The increase in BMD tests among men was estimated assuming that they should account for approximately 20% of all BMD tests (Table 36).
Using these estimated increases and decreases in BMD tests, the net budget impact was determined. The total cost of a BMD test is $106. (46) Table 37 summarizes the annual budget impact of shifting the utilization of BMD tests from areas of overuse to areas of underuse as estimated above. The annual cost avoidance due to a decrease in annual hip refracture rates was also included to estimate overall costs.
This economic analysis did not include the cost avoidance associated with a decrease in refracture rates in fractures other than hip fractures. Hip refractures are the most resource intensive types of fractures and therefore were the focus of this analysis.
The Ontario Schedule of Benefits (46) presently funds BMD testing every 24 months for low-risk persons. This review showed that for low-risk persons who have a normal baseline BMD (T score <10) and a normal rate of bone loss (≤ 1%/year) established through a repeat BMD at a 3-5 year interval, further testing is probably not necessary for another 7 years.
The Ontario Osteoporosis Strategy was announced by the Minister on February 2nd, 2005. The implementation of the strategy began soon thereafter. The Osteoporosis Strategy has five components aimed at health promotion and disease management.
The Strategy (Osteoporosis Action Plan) prepared by the Osteoporosis Action Plan Committee for the Ministry of Health and Long-Term Care and Osteoporosis Canada, identified potential gaps and inappropriate use of BMD and made the following recommendations:
In response to the above recommendations, the Ministry of Health and Lon-Term Care, through the Osteoporosis Strategy, has funded the following agencies to develop guidelines and requisition relating to BMD testing:
BMD results are influenced by the precision of the test that is, in turn, dependent on the equipment used, the standards adopted, and the skill of the technician conducting the measurements. In Ontario, at least 2 different standards for BMD testing are being used: standards of the International Society of Clinical Densitometry used mainly in hospitals, and standards of the Canadian Association of Radiologists, used by independent health facilities. Experts suggest that variability in results obtained from different test facilities may lead to unnecessary retesting.
There is presently no requirement for standardized equipment, procedure, or reporting format.
Presently, information relating to the use of BMD in Ontario can only be obtained from administrative databases that do not provide information on patient risks, test results, or patient outcomes. A registry for BMD tests would provide data for further assessment of BMD tests and form the basis for future standards and policy.
The increased use of BMD in Ontario since 1996 appears to be associated with an increased use of antiresorptive medication, and a decrease in hip and wrist fractures. Trends showed that BMD use has been moving in the right direction, since the growth in BMD use was mainly in people age 65 years and older. However, some areas for improvement were identified.
Future efforts to improve the appropriate use of BMD tests in Ontario need to focus on:
Some initiatives such as developing guidelines for men and perimenopausal women, and developing a standardized requisition form for BMD testing, are currently in progress under the Ontario Osteoporosis Strategy.
|Manufacturer||Device Name||License no.||Device Category|
|Norland, A Coopersurgical Company, Fort Atkinson, USA.||XR Series X-ray Bone Densitometer|
Excell Bone Densitometer
ExcellPlus Bone Densitometer
Apollo Bone Densitometers
|Hologic Inc., Bedford, MA, USA.||Sahara Clinical Bone sonometer|
Delphi Bone Densitometer
Discovery QDR series Bone Densitometer System
Explorer QDR Series X-ray Bone Densitometer
QDR 4000 Bone Densitometer
QDR 4500 Bone Densitometer
|GE Medical Systems Ultrasound and Primary Care Diagnostics, LLC.|
Madison, WI, USA.
|DPX Bone Densitometer|
DPX Bravo Bone Densitometer
DPX Duo Densitometer
Lunar IDXA Bone Densitometer
|Osteometer Meditech, Inc. Hawthorne, CA, USA.||X-ray Bone Densitometer|
Dexacare G4 X-ray Bone Densitometer
|Strategic Medizintechnik GMBH, Pforzheim, DE||XCT 2000 X-ray Bone Densitometer|
XCT 3000 X-ray Bone Densitometer
960 X-ray Bone Densitometer
|Recommendation by||Year||BMD measurement recommended for||Recommend method||Frequency|
|Ontario Osteoporosis Strategy (178)||2003||Ontario Guidelines for the Prevention and Treatment of Osteoporosis does not address BMD testing.|
Ontario Health Insurance Plan
|DXA at the hip and/or spine||Annual for people at high-risk of osteoporosis and once every 2 years for people at low risk|
|Canadian Task Force on Preventive Health Care (44)||2004||Postmenopausal women (Grade B) who (a) are ≥ 65 yrs old (b) <60 kg (c) have history of previous fracture, (d) have an ORAL score ≥9, or (e) have a score ≥6 on the SCORE questionnaire (grade B) Insufficient evidence to recommend using bone turnover markers to predict fracture||Use SCORE questionnaire or ORAL instrument to predict low BMD (grade A)|
BMD screening using DEXA to prevent fractures in post menopausal women with a risk factor (Grade I)
|Osteoporosis Canada (Former Osteoporosis Society of Canada) (45)||2002||BMD measurement for men or women age >65 (Grade A)|
-Targeted case-finding for those with increased risk (1 major or 2 minor)* adults age 50 – 65 yrs
|Hip or spine DXA the most accurate tool. Access to BMD measurement should not be limited by decision tools based on clinical risk factors (Grade A) Quantitative ultrasound may be considered for diagnosis of osteoporosis but not for follow-up at this time.|
Bone turnover markers should not yet be used for routine clinical management.
|-Monitor using central DXA in 1–2 years after initiating therapy|
-monitor height loss with thoracolumbar spine X-ray
|Manitoba (179)||2000||Manitoba Bone Density Program Targeted testing for:|
-Vertebral or nonvertebral fragility fractures proven by x-ray.
-Osteopenia or osteoporosis proven by x-ray
-Systemic corticosteroid therapy>3 months/year
-Prolonged amenorrhea prior to age 45 years if results needed to decide on hormonal or drug therapy
-Women>age 65 years if results needed to decide on hormonal or drug therapy
|Follow-up of previous BMD (initial recommended interval 3 years for most patients, 1 year for patients on systemic corticosteroid therapy|
|BC Health Services (43)||2005||DXA|
Risk factors same as those identified by Osteoporosis Canada.
Did not recommend screening for women <65 or as part of routine evaluation around the time of menopause.
|Follow-up BMD measurements not considered necessary prior to 2 yrs after previous measurement except in people on high dose of prednisone for >/= 3 months or with existing fractures with very low bone density|
|US Preventive Service Task Force (180)||2002||-Women ≥ 65 years|
-Women ≥ 60 years & at increased risk for osteoporosis
-Postmenopausal women <60 or between 60–64 yrs not at increased risk: no
recommendation for or against screening
|Number needed to screen to prevent 1 hip fracture in 5 years = approximately 1,000 or less|
|International Society for Clinical Densitometry 2004 (18)||2004||-Women≥ 65 years-|
Postmenopausal women <65 yrs with risk factors
-Men ≥ 70 yrs
-Adults with fragility fracture or disease associated with low bone mass or bone loss.
-Adults taking medication associated with low bone mass or bone loss
-People receiving treatment or in whom evidence of bone loss would lead to treatment
|DXA @ posterior-anterior spine & hip & @ forearm if spine or hip not feasible.|
OHIP Codes for BMD testing:
Low-risk patient X152, × 153
High-risk patient X149, X155
Codes for determining fractures:
The data sources for this analysis will include the National Ambulatory Care Reporting System (NACRS) for visits to the emergency department to identify fractures and Canadian Institute for Health Information (CIHI) hospital claims databases to exclude fractures related to malignancies and epilepsy.
In NACRS data
Transport accidents (V01-V99)
In NACRS data look in diagnosis code fields 1 to 3 to identify fractures and in DAD most responsible diagnosis. Identify fracture patients in NACRS and DAD and if in both then use DAD for assigning fracture since required a hospitalization and expect coding to be better than in ED.
|S72||Fracture of femur|
|S72.0|| Fracture of neck of femur|
Fracture of hip NOS
|S72.1|| Pertrochanteric fracture|
|S72.3||Fracture of shaft of femur|
|S72.4||Fracture of lower end of femur|
|S72.7||Multiple fractures of femur|
|S72.8||Fractures of other parts of femur|
|S72.9||Fracture of femur, part unspecified|
|Fracture of forearm|
|S52.0||Fracture of upper end of ulna|
|S52.1||Fracture of upper end of radius|
|S52.2||Fracture of shaft of ulna|
|S52.3||Fracture of shaft of radius|
|S52.4||Fracture of shafts of both ulna and radius|
|S52.5||Fracture of lower end of radius|
|S52.6||Fracture of lower end of both ulna and radius|
|S52.7||Multiple fractures of forearm|
|S62||Fracture at wrist and hand level|
|S62.0||Fracture of navicular [scaphoid] bone of hand|
|S62.1||Fracture of other carpal bone(s)|
|S62.2||Fracture of first metacarpal bone Bennett’s fracture|
|S62.3||Fracture of other metacarpal bone|
|S62.4||Multiple fractures of metacarpal bones|
|S62.8||Fracture of other and unspecified parts of wrist and hand|
|Johnell et al., 2004 (181)|
|Longitudinal study, Malmo, Sweden|
5 year follow-up
|N = 2,847 with low energy fractures @ spine, hip, & forearm||Poisson model to calculate age & sex-specific mortality rate & compare with that of general population||Higher mortality rate for men than women after hip, spine & shoulder fracture but rate relative to general population similar for both sexes. Rate decreased over 5 years (hip - from RR of 13 to RR of 4.3 for age 60). Mortality risk highest for spine and did not increase over general population risk for forearm fracture.|
|Pande et al., 2006 (11)|
|Prospective case control|
|N = 100 Consecutive men hospitalized with low trauma hip fracture. Mean age = 79.9 years Control n=100 matched without fracture, mean age 75 yrs||BMD measurement. Mortality from registers.|
Kaplan Meier survival curve analysis and a Cox proportional hazard model to determine factors for increased mortality.
|Significantly more patients with hip fracture had comorbid conditions and T-score<–2.5 (83% vs 39%) compared with control.|
Mortality @ 1 yr, 47% pts vs 1% for control. Mortality @ 2 year 63% for treatment group vs 12% in control. (log rank test62.6, P=.0001)
Most common causes of death: bronchopneumonia, heart failure, & ischemic heart disease
Factors associated with mortality after hip fracture: older age, residence in nursing/residential home before fracture, comorbid disease & poor functional activity before fracture.
Often disabled with poor quality of life. – 7% could not walk, 12% required residential accommodation lower QOL
|Barrett et al., 2003 (182)|
Male & female
|Case-control study||N = 81,181 Medicare recipient vs matched control with no fracture||Compared 90 day & 1-year mortality rate and pulmonary embolism rate||90 day mortality rate for entire US Medicare population after hip fracture was 13%, and after pelvic fracture was 9%. At 1-year after hip fracture, risk of death was 1.6 times that of matched control. Fractures of pelvis, nonhip, femur, & proximal humerus also associated with substantial mortality even a year after fracture. Death rates increased for age for both fracture cases & controls. For both men & women, relative risks of death (compared with controls) decreased with age for hip and pelvis fractures.|
|Jalava et al., 2003 (183)||N=677 women (84 men) and men with primary or secondary osteoporosis 352 had morphometric vertebral fracture||3.2 years follow-up||Mortality = 5.5%|
People with prevalent vertebral fracture had a 4.4 fold higher mortality rate compared with people with no prevalent fracture. After adjustment for medication, number of disease, use of oral corticosteroid, alcohol intake, serum albumin and erythrocyte sedimentation rate, renal function, height, weight, gender and age, the point estimate remained elevated but no longer statistically significant (HR 2.4 95% CI, 0.93–6.23).
|Holmberg et al., 2005 (154)||Longitudinal||22,444 men & 10,902 women identify incident hip fracture 137 women and 181 men had low energy hip fracture||Men 16 yrs|
Women 11 yrs
|Nonfracture population: mortality rate during follow-up 6.4% in women & 15% in men. Mean age of death 61 years. In hip fracture population: mortality rate was 16.8% in women @ average of 2.5 years. & 40.5% in men at average of 3.25 yrs follow-up. Mean age of death 64 yrs for women & 66.4 yrs for men.|
Database: Ovid MEDLINE(R) <1996 to August Week 3 2006>
Database: EMBASE <1980 to 2006 Week 34>
Search date: August 30, 2006
Databases searched: OVID MEDLINE, MEDLINE In-Process and Other Non-Indexed Citations,
EMBASE, INAHTA, Cochrane Library
Database: Ovid MEDLINE(R) <1996 to August Week 3 2006>
Database: EMBASE <1980 to 2006 Week 34>
Database: Ovid MEDLINE(R) <1996 to June Week 1 2006>
Database: EMBASE <1980 to 2006 Week 24>
Database: Ovid MEDLINE(R) <1996 to April Week 4 2006>
Database: EMBASE <1980 to 2006 Week 18>
|Melton et al., 2000 (Longitudinal study) (184)|
|Prospective population-based cohort study (348 men & 351 women)|
N= 351 women
|Age 21-93 yrs||DXA BMC & areal BMD for total body, L-spine (2-4), proximal femur, & forearm @ baseline, 1, 2, & 4 yrs. Rate of bone loss cross-sectional vs longitudinal, for ≥70 yrs vs <70 yrs||Cross-sectional data overestimated Rate of bone loss|
Rate of bone loss greater for ≥70 yrs than for <70 yrs group; may be due to younger women on HRT
|Chapurlat et al., 2000 (57) OFELY|
(Longitudinal study - 3 years) France
|Prospective population-based cohort study|
N = 272
|Premenopausal Perimenopausal Age 31-59 yrs healthy Caucasian randomly selected Excluded diseases or medications affecting bone, pregnancy||Annual DXA BMD total body, L-spine & hip, serum estrogen, osteocalcin, procollagen peptide, bone alkaline phosphatase, urine biomarkers for bone resorption or formation||Perimenopausal – no significant bone loss, small but significant increase @ FN, total hip & L-spine.|
Perimenopausal: rapid & diffuse bone loss related to decreased estrogen secretion (–0.1%/yr to –0.6%/yr)
|Bainbridge et al., 2004 US (185)|
(part of Michigan Bone Health Study)
|Prospective population-based cohort|
N = 614 Caucasian women
6 years follow-up
|Pre, peri & post menopausal community-dwelling (age 24-50yrs) from sampling lists.||Annual DXA BMD @ FN & L-spine (2-4), info on BMI, reproductive history, menstrual status, diet, lifestyle & medical history Linear regression analysis for association between risk factors & rate of bone loss|
|Goulding et al., 1999 (59)|
Prospective cohort 1 year
|Prospective cohort study|
N = 80
1 year follow-up
|Healthy women 40–79 years|
No previous test for lactose intolerance Excluded: therapy or conditions affecting bone e.g., GI surgery, radiotherapy, hyperthyroidism, hyperthyroidism etc.
|Effect of aging on lactose malabsorption & determine Ca intake & BMD loss Test for lactose malabsorption Baseline & @ 1 year: DXA BMD @ radius hip, spine & total body. Urinary biomarkers for bone formation & bone resorption||Ca intake only reduced in malabsorber 70 – 79 yrs|
BMD change in 1 year:
|Sellmeyer et al., 2001 (186)|
Prospective cohort (7 yrs)
|Prospective cohort study|
N = 1,035
|Caucasian women>60 yrs|
Mean age 73–74 years from population listings (no exclusion criteriastated)
|Animal protein intake from food frequency questionnaire, DXA BMD @ total hip & subregions @ baseline, @ year 2, & mean of 3.6 years later. Hip fractures assessed for 7+/-1.5 years. Linear regression analysis||High Animal/vegetable protein ratio bone loss @ femoral neck|
=0.78%/yr vs 0.21%/yr with a low ratio
Hip fracture higher with a high ratio (RR = 2.7, P=0.04)
|Chittacharoen et al., 1997 (58)|
50 surgical menopausal women without hormone replacement vs perimenopausal
N = 50 surgical (S) menopausal women & 50 controls (perimenopausalwomen)
N = 50
S menopause < 9 yrs
S. menopause > 9 yrs
|Study patients: women who had surgical & menopause and did not receive hormone replacement||Both study & control groups had DSA @ lumbar spine, femoral neck, total body, distal radius & midradius. Body, height, weight & BMD also assessed. Rate & pattern of bone loss were compared. For the surgical menopausal group was stratified according to postmenopausal period less than9 ears vs longer than 9 years.||BMD were significantly lower for the surgical menopausal women @ all skeletal sites measured. With postmenopausal period less than 9 years, the rate of bone loss was higher @ the lumbar spine & distal radius while for postmenopausal period longer than 9 years, the rate of bone loss was higher at the femoral neck compared to bone loss @ other sites.|
|N/Mean age||Rate of BMD change in Spine (% per year)||Rate of BMD change in Femoral Neck (% per year)||Rate of BMD change in Total hip (% per year)|
|Melton et al., 2000 (184)|
|N= 351 women|
<70 years old
≥70 yrs old
0.32 (lateral spine)
|Chapurlat et al., 2000 (57) OFELY (Longitudinal study - 3 years)|
|N = 272|
|Bainbridge et al., 2004 (185)|
(Prospective cohort 6 years)
|N = 614|
|(%/yr of T-score)|
|Goulding et al., 1999 (59)|
|N = 80|
|Sellmeyer et al., 2001|
Prospective cohort (7 years)
|N = 1,035 Caucasian|
|High animal protein|
Low animal protein
|Chittacharoen et al., 1997|
50 surgical menopausal women without hormone replacement vs perimenopausal
|N = 50|
S menopause < 9
S.menopause > 9
2.70 (distal radius)
|Rate of BMD change in Women in the Placebo Arm of RCTs on Antiresorptive Therapy|
|Montessori et al., 1997*(80)||Postmenopausal osteopenic women Mean age 62.9 yrs Mean lumber BMD 0.675 (Z-score< –1)on calcium alone||0.17 (95% CI –1.56; 1.90) in 3 years|
Average (+0.06 per year)
|–2.97 (95% CI-4.75; –1.19) in 3 years|
Average –0.96 per year, maximum – 1.6 per year
|Wasnich et al., 1999(187)||Early post-menopausal Caucasian women on Placebo, T-score< –2||– 1.9 (in 2 years) Average –0.95 per year||NR||–2 (in 2 years)|
|% BMD Changes From Baseline in Women in the Placebo Arm of RCTs on Antiresorptive Therapy|
|Bone et al., 1997|
women age 60–85
yrs, T-score< –2
|Osteoporotic elderly women on placebo Mean age 71.1 yrs Mean spine BMD 0.71 (0.09)g/cm2||0.56 (SE 0.44) in 2 years|
|-1.51 (SE 0.58) in 2 years|
(average –0.75 per year)
|McClung et al., 2004|
postmenopausal women (5mg) Mean age 51–55 years)
Mean age 53.7 year
Mean L-spine BMD
|–3.2(SE 0.34) in 6|
(average – 0.53% per year)
|-3.5 (SE 0.32) in 6 year|
(average –0.6% per year)
|-2.3 (SE 0.26) in 6 years|
(Average –0.4% per year)
|Cummings et al. 1998 (189)||Placebo|
Mean age= 67.7 (6.1)
Mean BMD (spine)
0.842 (0.13). FN 0.593 (0.06) L-spine
|+1.5 in 4 years (Average +0.37% per year)||-0.8 in 4 years (Average –0.2% per year)||–1.6 in 4 years Average –0.4% per year)|
|Bell et al., 2002 (190)|
|African American Women, mean age 65.9 yrs, mean L-spine BMD 0.80 (0.01)g/cm2 on placebo||+0.9 (SE 0.6) in 2 years|
Average +0.45 per year
|+0.5 (SE 1.1) in 2 years|
Average +0.25 per year
|–1.1 (SE 0.7) in 2 years|
Average –0.55 per year
|Pols et al., 1999 (191)||Women postmenopausal>3 yrs, <85years old Placebo|
Mean spine BMD =0.72 (0.08) mg/cm2
|0.1 (SD 3.4) in 1 year||–0.2 (SD 4.5) in 1 year||0.1 (SD 3.0) in 1 year|
|Hooper et al., 2005 (192)||Postmenopausal women, mean age 52.6 yrs|
Mean spine T-score – 0.432 on placebo
|–2.5 in 2 years Average –1.25 per year||–2.5 in(2 years Average –1.25 per year||–1.8 in 2 years Average –0.9 per year|
|Clemmesen et al.,|
|Postmenopausal women mean age = 70 years, mean spine BMD 0.747 g/cm2 on placebo||1.7 in 3 years Average 0.6 per year||–2.6 in 3 years Average –0.86||Trochanter –0.04 in 3 year Average –0.013 per year|
|Harris et al., 1999 (194)|
postmenopausal women 1<85 yrs >/=1 vertebral fracture
|Postmenopausal women <85 years with ≥ 1 vertebral fracture||1.1 In 3 years Average 0.33 per year||–1.2 in 3 years Average –0.4 per year||Trochanter –0.7 in 3 years Average –0.23 per year|
|Rate of BMD change Spine % per year (SD)||Rate of BMD change Femoral Neck % per year(SD)||Rate of BMD change Total hip % per year|
|Melton et al., 2000 (195) (Longitudinal) US||N=348 men|
|Van Pottelbergh et al., 2003 (196) Belgium||N = 214|
Age 71–86 yrs
|Knoke et al., 2003 (150)|
(Rancho Bernardo Study) US
|N = 1,214|
Mean = 70.6 yrs
|–0.5 (men & women)|
29% of men lost at least
1%/yr (especially in people with >1%/yr weight loss
|Bakhireva et al., 2004|
Age 45–92 yrs
|Cauley et al., 2005|
|N = 5,995|
|+7% for every 5 yrs increase in age (Average +1.4%/yr)||–2.6% every 5 yrs increase in age (Average –0.52%/yr)|
|Ensrud et al., 2005 (60)|
|N = 1,342|
Age >65 yrs
|Weight gain 0.1|
Stable weight –0.3
Weight loss>5% –1.4 r
|Naves et al., 2005 (198) Spain||N = 308|
Age ≥50 yrs
|–0.0011 (.009) g/cm2/yr||0.0008 (.011) g/cm2/yr|
|% BMD Change From Baseline in Men in the Placebo arms of Randomized Controlled Trials*|
|Gonelli et al., 2003 (119) †|
|Men with primary osteoporosis treated with placebo & calcium alone Mean age = 56.6 years, mean FN BMD 0.622 g/cm2||–1.2 (3 years)|
Average –0.4 per year
|–1.2 (3 years)|
Average –0.4 per year
|–0.3 (3 years)|
Average –0.1 per year
|Ringe et al., 2004 (120) †|
Open label RCT
|Men with primary osteoporosis on alfacalcidol, mean age = 53.3|
Mean FN T-score = –2.56
|3.5 in 3 years|
Average 1.16 per year
|2.3 in 3 years|
Average 0.76 per year
|% Change/year in Total Hip Bone Mineral Density (95% Confidence Interval)|
|Gained >5% weight||Weight Stable||Lost >5% weight||P-value test for trend|
|Entire cohort †||0.1 (–0.1, 0.4)||–0.3 (–0.4, –0.3)||–1.4 (–1.6, –1.2)||<.001|
|BMI<25 kg/m2||0.2 (–0.3, 0.7)||–0.4 (–0.5, –0.2)||–1.4 (–1.9, –0.9)||<.001|
|BMI 25.0–29.9 kg/m2||0.2 (-0.2, 0.5)||–0.4 (–0.5, –0.3)||–1.2 (–1.5, –0.9)||<.001|
|BMI >30 kg/m2||–0.2 (–0.8, 0.5)||–0.3 (–0.5, 0.0)||–1.7 (–2.2, –1.2)||<.001|
|BMI>30 kg/m2, trying to lose weight||0.5 (–0.3, 1.3)||–0.1 (–0.4, 0.1)||–1.7 (–2.4, –1.1)||<.001|
|Include only RCTs||Clearly defined question||Description of search strategy||Inclusion/exclusion criteria||Study selection & data abstraction by > 1 person||Assessed Quality of Studies||Method of meta-analysis described||Use Individual patient data in Meta-analysis|
|Wasnich et al., 2000 (64)||√||√||Based on review articles & abstracts||√||NR||NR||√||NR|
|Cummings et al., 2002 (65)||√||√||√||√||√||NR||√||No|
|Hochberg et al., 2002 (66)||√||√||NR||√||NR||NR||√||No|
|Guyatt et al., 2002 (63)||√||√||Not applicable||From meta-analysis||NR||√||√||Summary of meta-analysis|
|Delmas et al., 2004 (67)||√||√||No, based on studies from another meta-analysis||√||NR||NR||√||√|
|Watts et al., 2004 (68)||√||√||NR||√ †||NR||NR||√||√|
|Watts et al., 2005 (69)||√||√||NR||√ †||NR||NR||√||√|
|Wasnich et al., 2000 (64)||Hochberg et al., 2002 (66)||Delmas et al., 2004 (67)||Cummings et al., 2002 (65)||Watts et al., 2004 (68) & 2005 (69)|
|Meta-analysis/Study||Year of studies||Antiresorptive Drug (number of studies)||Type of BMD & Type of Fracture Studied||% change in BMD vs % change in fracture risk||Other Findings|
|Wasnich et al., 2000 (64)||13 Randomized placebo controlled trials †||Alendronate (4)|
|% Changes in spine BMD & hip BMD vs RR of vertebral fracture||The model predicts:|
-treatments that increase spine BMD by 8% would reduce vertebral fracture risk by 54% and change in BMD explained mot of the effect (41% risk reduction) Treatments that increased hip BMD by 5% would reduce vertebral fracture risk by 50% with 38% attributable to BMD
|Poisson regression:CI large for individual trials. Substantial variability inantifracture efficacy at anygiven level of change in BMD.|
Overall trials reportinglarger increase in BMDtended to have greaterreductions in vertebralfracture risk.
-A small but significant riskreduction of 20-22% withno measurable change inspine BMD.
|Hochberg et al., 2002 (66)||18 double-blind placebo controlled RCTs † (n = 26,494) 1–3.5 yrs follow-up||Etidronate (2)* Alendronate (7) Tilodronate (2) Estrogen (2) Risedronate (4) Calcitonin (3) Raloxifene (2)||Spine & Hip BMD, bone markers (BCM), & nonvertebral fractures||At 1 year: 6% in spine BMD (treatment vs placebo) was associated with a 39% reduction in nonvertebral fracture risk. 3% in hip BMD (over placebo) was associated with 46% risk reduction 70% decrease in bone resorption marker associated with 40% risk reduction||There was a significant association between the amount of increase in BMD @ the spine &hip at 1 year and in risk of incident nonvertebral fractures (P= .02 & .006) without an independent effect of treatment. Multiple regression line not perfectly linear but BMD or BCM appear to explain a significant part of the risk reduction. The association remained significant in sensitivity analysis. Changes in BMD were correlated with changes in bone markers (P</= .002)|
|Guyatt et al., 2002 (63)||Summary of 8 meta-analysis of placebo controlled RCTs *†||Calcium (15)|
Vitamin D (25)
|Relative risk of vertebral & nonvertebral fractures, absolute difference in event rates, changes in BMD, relationship between BMD changes & fracture risk reduction||Reduction in vertebral fracture risk:|
Alendronate, calcitonin about 50% Vitamin D, etidronate, risedronate, raloxifene about 33.3% Nonvertebral fracture: significant reduction only for Alendronate (50%) (At least 31%) and risedronate (33.3%) (At least 13%) & may be less in low-risk population.
|Based on Poisson regression analysis of individual trial data: Vertebral fracture relative risk reduction 20% with no effect on BMD and an additional 25% associated with BMD.|
(BMD explained 25/45% of risk reduction)Nonvertebral fracture risk reduction: No significant relationship between BMD and risk reduction. BMD is not helpful for predicting the impact of antiosteoporosis treatment on nonvertebral fractures.
|Dalmas et al., 2004 (67)||16 placebo-controlled RCTs†||Same as Hochberg with omission of 3 studies to correct for discrepancies in reported BMD & person years||Spine & Hip BMD, & nonvertebral fractures||No association between the extent of reduction in nonvertebral fracture risk and increases in BMD at the spine or hip at one year or at study end point. Larger increase in BMD @ 1 year were not associated with greater reduction in nonvertebral fracture risk (P = .12) for L spine & P=.11 for hip). Larger increases in BMD from baseline to study endpoint were not associated with greater reduction in nonvertebral fracture risk (P=.47 for L spine and .60 for hip).|
|Cummings et al., 2002 (65)||12 blinded placebo-controlled RCTs † 1990-2000 (n = 21,404)||Etidronate (2) Alendronate (3) Tiludronate (1) Estradiol (1) Risedronate (2) Raloxifene (1) Calcitonin (2) (Vit D & Calcium for both groups)||DXA Spine BMD & Vertebral fractures||1% spine BMD (Treatment vs placebo) was associated with 0.03 in RR of spine fractures (95% CI, 0.02-0.05, P=.002)||Based on in BMD, expected to reduce fracture risk by 20% (RR= 0.8), but treatment reduced risk of vertebral fracture by 45% Alendronate: 3.9% in BMD @ 1 year explained 16%of the (47%) in risk of vertebral fracture|
|Watts et al., 2004 (68)||3 pivotal double-blind parallel placebo controlled RCTs (VERT & HIP) (n=3,224) Meta-analysis using individual patient data||Risedronate 2.5 mg or 5mg (n = 2,047) or placebo (n=1,177)047) For up to 3 years + 1,000 mg Calcium & Vit D supplement (if serum 25 OH vitamin D<40 mol/L)||DXA spine & femoral neck BMD @ baseline and 6-month intervals. Lateral thoracolumbar (T4–L4) radiographs were @ baseline & annually for 3 years to identify vertebral fractures||Incidence of vertebral fracture: 15% for spine BMD<0%; 9.5% for BMD = 0 - <5%;|
10.2% for BMD >5%
Changes in spine BMD explained 18% of treatment effect & femoral neck BMD explained 11% of treatment effect
|Risk reduction similar for BMD <5% & BMD >5% (49% vs 41%, P=.77) Patients showing an increase in BMD had lower fracture risk than patients showing a decrease in BMD, greater increases in BMD did not necessarily predict greater reduction in fracture risk.|
|Watts et al., 2005 (69)||3 pivotal double-blind parallel place bo-controlled RCTs (VERT) (n=3,979)|
Meta-analysis using individual patient data
|Risedronate 2.5 mg or 5mg or place bo For up to 3 years + 1,000 mg Calcium & Vit D supplement (if serum 25 OH vitamin D<40 mol/L)||3DXA L-Spine & femoral neck BMD @ baseline & 6-month intervals and Radiologically confirmed nonvertebral fractures||Risedronate risk fracture by 32% (HR 0.68, 95% CI,, 0.54-0.85, P< .001). Changes in L-spine BMD explained 12% and change in FN BMD explained 7% of effect of risedronate on nonvertebral fracture incidence.||Similar incidence of nonvertebral fracture whether there is an increase or decrease in BMD with treatment (7.8% for pts whose spine BMD decreased vs 6.4% in patients with increased spine BMD. Similar for FN BMD (Incidence 7.6% for pts with increased FN BMD vs 7.5 for people with increased FN BMD) Changes in BMD as measured by DXA did not predict the degree of reduction in nonvertebral fracture.|
|Meta-analysis/Study||Year of studies||Anti-resorptive Drug||Type of BMD & Type of Fracture Studied||% change in BMD vs % change in fracture risk||Other Findings|
|Sarkar et al., 2002 (70)|
N = 7,705
postmenopausal women with osteoporosis
|Randomized to Raloxifene (60 or 120 mg/day) or placebo|
Raloxifene patients pooled for logistic regression analysis
|DXA lumbar spine or femoral neck BMD @ baseline & annually ANND vertebral fractures||@ 3 years, raloxifene group had a 36% lower risk of vertebral fracture compared with placebo group.||Women with the lowest baseline lumber spine or femoral neck BMD had the greatest risk for femoral fractures. % change in BMD accounted for 4% of the observed vertebral fracture risk reduction, and the other 96% of the risk reduction remains unexplained.|
|Chapurlat et al., 2005 (71)|
Fracture Intervention Trial (FIT)
|Multicenter placebo controlled RCT.|
N = 5,383 women With low BMD Randomly assigned to either alendronate or placebo (post hocanalysis)
|Alendronate: 5mg/D 1st 2 yrs then increased to 10 mg|
If calcium intake<1,000 mg, also receive calcium & 250 IU vitamin D supplement Alendronate & Placebo patients each divided into 4 categories BMD loss >4%, 0-4%, gain 0-4%, >4%
|Baseline & annual L-spine & hip BMD with QDR & Vertebral fractures Lateral spine radiographs obtained at baseline and @ end of 3 year follow-up for FIT & 4 years for FIT-II.||Lost: -spine BMD Alendronate 10% Placebo 40% Lost Hip BMD Alendronate 19% Placebo 47%|
Among patients who adhere to treatment with alendronate, even those who lose BMD benefit from a substantial reduction in risk of vertebral fracture.
|@ 1 year:|
Reduction in risk of vertebral fracture, alendronate vs placebo: Lost 0%-4% L-spine BMD
OR 0.4 (95% CI, 0.16–0.99)
Lost >4% L-spine BMD OR 0.15(95% CI, 0.02–1.29)
Gained 0%–4% spineBMD OR 0.49 (95% CI, 0.3–0.78)
Gained>4% spine BMD OR 0.46 (0.32–0.66)
Lost 0-4% hip BMD OR 0.47 (0.27–0.81)
Lost >4% hip BMD OR 0.61 (0.11–3.45)
Gained 0%–4% hipBMD OR 0.49 (0.34_0.71)
Gained >4% OR 0.34 (0.18–0.62) No benefit fromalendronate for people who lost BMD@ bothhip & spine
|Cummings et al., 2005|
FIT Trial MORE Trial
Multicenter placebo controlled
N = 2,634
|FIT: women completed 2 yrs of treatment with alendronate (5 mg/day)|
MORE: women completed 2 yrs of treatment with raloxifene (60 or 120 mg/day)
|changes in L-spine and total hip BMD in year 1 compared with year 2||FIT: 92% of pts who had the >4% loss in hip BMD in year 1 gained BMD in year 2 (92%), & gained the most: average 4.8%.|
Patients who had the greatest gain in hip BMD (>8%) in year one were least likely to gain BMD in year 2 (36%). Average, lost 1% BMD. Similar pattern for spine BMD and in the placebo group
|MORE: Women who lost >4% FN BMD during year 1 of raloxifene had a 79% chance of gaining BMD in year 2, with an average gain of 4.0%. Patients who gained>8% in yr 1 had a 22% of gaining BMD in year 2 and on average lost 2.8% in yr 2.|
Concluded that most women who lost BMD during the first year of treatment with alendronate or raloxifene will gain BMD if the same treatment is continued for a second year, illustrating the principle of regression to the mean & that effective treatment for osteoporosis should not be changed because of loss of BMD during the first year of use.
|Study||Sample size||Clear Inclusion/exclusion criteria||Method of randomization||Concealment||Blinding||Power calculation||% complete follow-up||Intention to treat analysis|
|Lyritis et al., 1997||100||√||Not stated||No||No||Not stated||73%||Not stated|
|Montessori et al., 1997* (80)||80||√||√ by computer in blocks of 4||Not stated||No||Not stated||80%||√|
|Ishida et al., 2004||396||√||√||Not stated||√ end point evaluation||Not stated||9% withdrawn||√|
|Bone et al., 1997 (188) (multicenter)||994||√||Not stated||Not stated||√||Not stated||81% @ 3 yrs 51%@ 10 yrs||√ Primary evaluation|
|Pols et al., 1999 (191) (Multicenter)||1,908||√||Not stated||Not stated||√||√||88% study group 90% Placebo||√ Primary evaluation|
|Cummings et al., 1998 (189) (Multicenter)||4,432||√||√ in blocks of 10 by in blocks of 10 by computer generated codes||√||√ All blinded to treatment & BMD results||√||96%||√|
|Wasnich et al., 1999 (187) (EPIC)||262||√||Not stated||Not stated||Unclear||Not stated||79%||√ for treatment group|
|Bell et al., 2002 (190) (multicenter)||65||√||Not stated||Not stated||√||Not stated||72% study group 67% placebo||√ Primary analysis|
|McClung et al., 2004 (81) (multicenter) Prevention||529||√||Not stated||Not stated||Single blind||Not stated||71.3% @ 6 years||√ modified|
|Evio et al., 2004 (84) (Single center)||60||√||Not stated||Not stated||Single||Not stated||77%||Not stated|
|Sambrook et al., 2004 (199) (multicenter)||487||√||√ Computer generated random allocation||√||√||Not stated||Alendronate 88% Raloxifene 86%||√ modified|
|Black et al., 2003 (85) (Multicenter)||238||√||Not stated||Not stated||√||√||95% @ 1 year||√|
|Gonelli et al., 2003 (119) (Single center)||77||√||Not stated||Not stated||Open label||Not stated||86%||No|
|Ringe et al., 2004 (200) (Single center)||167||√||Not stated||No||Open label Radiologist blinded to allocation||Not stated||88%||√|
|Hooper et al., 2005 (192)||383||√||√ computer generated||Not stated||√||√ 90% power for 3% difference||77%||√|
|Clemmesen et al., 1997 (193)||132||√||Not stated||Not stated||√||Not stated||73% @ 2 years||√|
|Fogelman et al., 2000 (201) (Multicenter)||543||√||Not stated||Not stated||√||√||70% @ 3years 65.4%||√|
|Harris et al., 1999 (194)||2,458||√||√ computer generated||√||√||Not stated||57%||√|
|Reginster et al., 2000 (202)||1,226||√||Not stated||Not stated||√||√ 90% power to detect 40% fracture reduction||81% @ 1 year|
|Ringe et al., 2006 (86)||316||√||Not stated||Not stated||Open label||Not stated||100% @ 1 year||√|
|Ettinger et al., 1999(203)||7,705||√||Not stated||√||√||√||77%||√|
|Sambrook et al., 2004 (199) (multicenter) Morii et al., 2003 (204)||See study under alendronate 284||√||Not stated||√||√||Not stated||87%||√|
|Black et al., 2003 (85) Postmenopaus al||See previous table|
|Hodsman et al., 1997 (205)||217 (Placebo 53)||√||Not stated||Not stated||Double blind|
|Neer et al., 2001 (206)||1,637||√||Not stated||Not stated||√|
|Not stated||81% for radiographs||Not stated|
|Body et al., 2002 (207)||PTH (20 or 40 ug) vs placebo 146 (PTH vs alendronate)||√||Not stated||Not stated||Double blind|
|McClung et al., 2005 (208) (multicenter)||203|
(PTH vs alendronate)
|√||Not stated||Not stated||Double blind|
|Lane et al., 1998 (209)||51|
(PTH vs PTH+estrogen)
Computer generated table
|Not stated||Not stated||Not stated||94%||Not stated|
|Cosman et al., 2005 (210)||126|
Alendronate vs Alendronate + PTH
Bycomputer in blocks of 18
|Not stated||No blinding (patient & physician); blinded outcome assessment||Not stated||78%||√|
|Finkelstein et al., 2003 (211)||83 (men)|
Alendronate vs PTH vs both
|No blinding to treatment; blinded assessment of BMD||Not stated||76%||√|
Include those with at least 1 BMD test
|Kurland et al., 2000(212)||23 (men)|
PTH vs placebo
|√ Inclusion Exclusion not specific||√|
By computer in blocks of 4
|Not stated||Not clear||√|
|Study||Patient/Drug||Effect on BMD||Effect on fracture||Authors’ conclusion|
|Alendronate Sawka et al., 2005 (118)||Meta-analysis Of Orwell & Ringe|
placebo or vitamin D
|Using Bayesian random effects model|
OR vertebral fracture in alendronate treated men 0.44 (95% CI, 0.23–0.83) OR nonvertebral fracture 0.6 (95% CI, 0.29–1.44)
|Alendronate decreases risk of vertebral fractures in men with low bone mineral or fractures.|
Insufficient data for effect on non-vertebral fractures
|Orwoll et al., 2000 (213)|
Double blind RCT
|Men mean age 63 yrs with femoral neck T-score>/= –2 or L-spine T-score< –1|
(Alendronate 10 mg + Ca+vit D) vs (placebo+Ca+vit D) (N= 146/95)
Follow-up = 2 years
|BMD increase significantly ↑ 7.15 in lumbar spine & 2.5% in femoral neck in Alendronate vs ↑ of 1.8% in L-spine & 0.6% @ hip of placebo. BMD significantly higher in Alendronate group @ each site.||3% alendronate vs 13% control had >/=10 mm height loss|
Vertebral fractures 0.8% alendronate vs7.1% control (P = 0.02) Effect independentof age.
|Significantly lower bone marker level in alendronate group|
|Ringe et al., 2004 (200)|
Open label RCT
|Men with primary osteoporosis|
Alendronate 10 mg vs alfacalciferol
(N = 68/66)
Mean age 52.1/53.3 @ 3 years (58/60 completed treatment)
|BMD over baseline|
L-spine 11.5% alendronate vs 3.5 control (P = .0001)
Femoral neck 5.8% vs 2.3% (P = .0015) 87% of alendronate vs 46% of control group had increase in spine BMD>/=3%, 63% vs 33% had increase in hip BMD>/=3%
|Vertebral fractures occurred in 10.3% alendronate vs 24.2% control (P = .04)|
57% reduction in vertebral fracture risk Change in height: –7.1mm in alendronate vs 13.1mm in control (P = .03)
|No significant difference in nonvertebral fractures|
Both treatments well tolerated. Hypercalciuria reported in 15.1% control vs 4.4% alendronate (P = .04)
|Gonnelli et al., 2003 (119)|
|Primary osteoporosis Alendronate 10 mg+ Ca vs Ca alone (n=39/38)|
|Alendronate group significant increase in spine BMD in each of 3 yrs of follow-up 4.2-8.8% Increased total hip BMD only significant in year 3 (3.9%)||BMD at lumbar spine appear to be the best method for monitoring effect of alendronate on bone mass in osteoporotic men (> least significant change al each year).|
|Sato et al., 2005 (121)|
|Ambulatory men after stroke >/=65 yrs|
Risedronate 2.5 mg oral vs placebo (n=140/140)
|BMD Alendronate group +2.5% vs control –3.5% (P<.001) Serum Ca+ decreased & PTH & 1,25(0H)2 increased in risedronate group but stayed low in control||Number of fall similar|
Hip fracture 2/140 inrisedronate vs
10/140 control RR 0.19 (95% CI, 0.04–0.89)
NNT for hip fracture16 (9–32)
|Ringe et al., 2006 (122)|
Single centre open label RCT
|Men with primary & secondary osteoporosis (Risedronate 5 mg+1g Ca & vitamin D) vs (placebo)|
Mean age 55.8 yrs vs
|Increase in spine BMD Risedronate 4.7% vs 1% control (P<.0001) Total hip BMD +2.7|
Femoral neck 1.8% vs 0.2% (P <.0001)
|New vertebral fractures risedronate 5.1% vs 12.7% control (P = .028)|
No significant difference in non-vertebral fracture risk.
Height change 1.1mm risedronate vs –4.6mm control
|Improvement in back pain greater in risedronate than control (P < .0001)|
|Study||Site||Mean % Change in BMD From Baseline (95% Confidence Interval During Treatment|
|After 1 year||After 2 years (SE when reported)||After 3 years|
|Lyritis et al., 1997 (214)|
90 day cyclical (400 mg × 20 days)
|Montessori et al., 1997 * (80)||Spine|
|4.25 (95% CI, 2.90; 5.59)|
2.73 (95% CI, 1.51; 3.95)
|5.67 (4.04; 7.29)|
1.44 (–0.60; 3.47)
|Ishida et al., 2004 (215)|
Postmenopausal osteoporotic women, age
50–75 years 200mg x2 weeks then none for 10 weeks
|Distal radius||– 0.5%|
|Bone et al., 1997 (188)|
women age 60–85 yrs, T-score< –2 (5mg) *
|Wasnich et al., 1999 (187)|
Early Caucasian postmenopausal (5mg)
|McClung et al., 2004 † (81)|
Early postmenopausal women (5mg) Mean age 51–55 years)
|Iwamoto et al., 2005 (216)|
Women 55–86 yrs old
BMD<70% of young adult mean (5mg)
|Gonelli et al., 2003 (119)|
Men with primary
Osteoporosis (5 mg)
|Evio et al., 2004 (84)|
women age 65–80 yrs (mean 71), T-score≤ –2.5
(10 mg+/- HRT)
|Sambrook et al., 2004 (217)|
Mean age 62 yrs, T-score ≤–2
|Cummings et al. 1998 (189)|
Women age 55–80 yrs
BMD @ femoral neck ≤ 0.68 g/cm2 (T-score ≤ –2) (5mg increased to 10 mg/d)
|Bell et al., 2002 (190) African-American women (45–88 yrs), T-score≤ –1.75, Alendronate(10 mg)||L-spine|
|Pols et al., 1999|
|Black et al., 2003|
T-score < –2.5 or ≥ –2 + risk factors
mg/day or PTH 100
ug or both
|Ringe et al., 2004 *|
(200) Men with
primary osteoporosis (10 mg)
|Hooper et al., 2005 (192)|
Women, early postmenopausal
|Clemmesen et al.,|
postmenopausal >1yr, 1–4 vertebral fractures (2.5mg)
|Harris et al., 1999 * (194)|
postmenopausal women l<85 yrs >/=1
|Reginster et al., 2000 (202)|
Women>5 yrs postmenopausal <85 yrs >/=2 vertebral fractures (5mg)
|Ringe et al., 2006 (122) Men (primary|
|Sarkar et al., 2002 (70)||L spine|
|Ettinger et al., 1999 (203)|
MORE >/=2 yrs postmenopausal women & 31–80 yrs old (120mg)
|Sambrook et al., 2004, (199) Mean age 62 yrs, postmenopausal|
|Morii, 2003 (204) (60mg)|
Japanese women osteoporotic, PM
|Black et al., 2003 (85)|
T-score< –2.5 or –2+ a major risk
|Hodsman et al., 1997 (205)|
5.1 (75 ug)
7.8 (100 ug)
0.5 (100 ug)
|Neer et al., 2001 (206) Postmenopausal & history of fracture||L-spine|
13.7 (40 ug)
5.1 (40 ug)
|Body et al., 2002 (207)(Teriparatide)|
|McClung et al., 2005 (208)|
Osteoporosis (20 ug)
|Lane et al., 1998 (209)|
Postmenopausal + cortisone induced- osteoporosis
(daily or cyclic PTH 25ug+estrogen)
|Cosman et al., 2005 (210)|
Women T-score <–2.5 or –2+fracture
(25 ug) PTH + alendronate
|Finkelstein et al., 2003 (211)|
Men T-score< –2
|Kurland et al., 2000 (212)|
Men T-score <–2.5 or Z-score< –2
|Hajcsar et al., 2000 (134)||Papaioannou et al., 2004 (87)||Khan et al., 2001 (218)||Juby et al., 2002 (219)||Khandwala et al., 2005 (132)||Vanasse et al., 2005 (88)|
|Study design||Retrospective analysis||Retrospective analysis||Retrospective analysis||Prospective cohort study||Retrospective review of medicalrecords||Retrospective population-based cohort study|
|Setting||3 Ontario Community hospital fracture clinic||4 tertiary care hospitals Hamilton, Ontario||1 tertiary care hospital, Edmonton, Alberta||Tertiary hospital, Edmonton, Alberta- pts from seniors’ clinic & a day program||1 hospital in Saskatoon, Saskatchewan||Province of Quebec – data from Quebec Health Insurance Board|
Complete follow-up Age, years
|N=108 (89% women)|
Representing 56.1% of patients with fragility fractures 108
|All patients with hip fracture|
|N = 156 (83% women)|
72% of patients with fragility fracture of distal radius/ulna
N = 112 (72%)
Mean 64 (range 41–91 years)
|N = 145 (73% women)|
All patients witha new hipfracture
145 complete questionnaire >65, Mean 72 & 77.7
|N = 174|
Admitted with fragility hip fracture
Mean age 82.5 (9.8) yrs
|N = 25,852 (77% women) with a fragility fracture (vertebral, hip, wrist, or humerus) in 1999 & 2000 25,852|
≥ age 65
|History of prior fracture %||39.8||39.2||16||22|
|Time after fracture, years||1||1||0.5–3||Acute & after rehabilitation||@ discharge|
|BMD testing, %|
Clinical diagnosis %
Vitamin D %
specific therapy %
22.2 (after index fracture)
BMD or clinical 50
61.6 (Ca or Vitamin D)
37.5 (HRT or bisphosphonate)
|Overall = 26%|
29% of women
18.3 (3% in men)
Evaluation recommended in
Recommended in 4% implemented in 3%
No significant difference in intervention rates by sex & history of fractures.
Regional BMD range 0–16%
Use of BMD with distance from BMD facility
|Study||Design||Population||Prevalent fractures||Incident fractures|
|Klotzbuecher et al., 2000 (94)||Meta-analysis 33 studies 1996–1999||peri/postmenopausalwomen||Wrist, vertebral, hip, pooled||Vertebral, hip, wrist, pooled|
|Kanis et al., 2004 (95)||Meta-analysis(11 large population–based studies 1994–2003||15,259 men and 44,902 women||Prior fractures||Hip, any fracture, osteoporotic fracture (with and without effect of BMD)|
|Haentjens et al., 2003 (96)||Meta-analysis of 9 cohort studies 1982–2001||Colle’s fractures & spine fractures||Hip fractures|
|Johnell et al., 2004 (98)|
Immediately following fracture
|Prospective longitudinal cohort study in Sweden 5-year follow-up||1,918 men and women identified by radiology to have a fracture at the spine, hip or shoulder||Spine|
|Relative risk of Hip, spine, and forearm over time stratified by age 60 & age 80 years|
|Papaioannou et al., 2005 (19)|
|Prospective multi-site population based Canadian cohort study 3-year follow-up||5,143 postmenopausal women who participated in the Canadian multicenter Osteoporosis Study mean age of group 66.4 (SD 9.6) –74.4 (SD 10.0) years||Vertebral Forearm Nonvertebral||Vertebral|
Main nonvertebral (wrist, hip, humerus, pelvis or rib)
Any nonvertebral fractures
|Bensen et al., 2005 (26) (Canada)||Analysis of prospective multisite Canadian CANDOO database||3,426 postmenopausal women registered in the CANDOO||Previous fractures after age 50 years||OR for vertebral fractures, hip fractures, wrist fracture, & rib fracture|
|Schousboe et al., 2005 (102)||Study of Osteoporotic Fractures (SOF) - Prospective cohort study in the US Mean follow-up 3.7 years||9,704 elderly community dwelling women, mean age 73.2 years (with wrist fracture) & 71.5 (no wrist fractures)||Previous wrist fractures since age 50 years||Hip fractures|
Radiographic vertebral fracture
|Van der Klift et al., 2004 (103)||Prospective population-based cohort study – Part of the Rotterdam Study Mean follow-up 6.3 years||4,216 men & women (2467 women) age>55 years mean age for subgroups 65.2– 68.6 years||Vertebral fracture||Vertebral fracture|
|Porthouse et al., 2004 (104)||UK comprehensive cohort study with anested randomized controlled trial on hipprotectors 2 year follow-up||4,292 women ≥ 70 years|
Mean age 76.9
|Previous fracture||Hip, nonvertebral, wrist fractures|
|Taylor et al., 2004 (24)||Study of Osteoporotic Fractures (SOF) - Prospective cohort study in the US||6,787 community-dwelling, ambulatory Caucasian women ≥age 65 (mean age 73.3 (SD 4.9) years from SOF with complete data||Any previous fractures since age 50 years||Hip fractures|
|Colon-Emeric et al., 2003 (105)||Analysis of data from the Baltimore Hip Studies and the Established Populations for Epidemiologic Studies of the Elderly (EPESE)|
Mean follow-up 6.0 and 1.6 years respectively
|Baltimore study: 549 men & women >/= 65 years of age with acute hip fracture (Mean age 80.9 (SD 7.4) years EPESE: 10,680 community-dwelling men & women age >/= 65 years. Mean age 73.8 (SD 6.7) years||Hip fracture||Hazard ratio for subsequent nonhip skeletal fracture|
|Naves et al., 2003|
|Prospective cohort study – Spanish cohort of the EVOS study|
Follow-up 8 years
|316 women and 308 men age> 50 years randomly selected from the EVOS cohort.|
Mean age 65 (SD 9) for men and women
|Prevalent vertebral fracture Prevalent and Incident vertebral fracture|
Intraobserver agreement = 92%, interobserver agreement of 90%
|Hip, Colles’, vertebral|
|Albrand et al., 2003 (OFELY)|
|Longitudinal cohort study of healthy ambulatory Caucasian volunteers in Rhone district of France, followed for a mean of 5.3+1.1 years||672 postmenopausal healthy ambulatory Caucasian women (mean age 59.1 years (SD 9.8 years)||All prevalent fractures after age 45 years||Fragility fractures|
|Pongchaiyakul et al., 2005 (99)||Part of ongoing Dubbo Osteoporosis Epidemiology study (DOES) – longitudinal, population-based study of risk factors for fracture & mortality in Australia (5 year follow-up)||114 men and 186 women (age> 60 years & free of illnesses that affect bone metabolism) randomly selected from the DOES database Mean age 69.8 years with vertebral deformity) & 69.4 years with no vertebral deformity||Asymptomatic vertebral deformity (at least –3SD in vertebral height) confirmed on radiograph||Any fracture|
Hip fracture Vertebral fracture Colles’ fracture Major fractures (major upper or lower limb and/or rib fractures)
|Hasseius et al., 2003 EVOS (101) Longitudinal||European Vertebral Osteoporosis Study – multicenter study to evaluate vertebral deformity – men and women followed for 10 years||Men & women age 50–80 years|
213 men (mean age 63 years) and 257 women (mean age 64 years)
|Vertebral deformity (–3 SD or –5 SD in vertebral height||Any incident fracture|
Any fragility fracture
|Szulc et al., 2005 (97)||A prospective study of osteoporosis and of its determinants in men (MINOS) in France Follow-up 7.5 years||791 men aged 51-85 years were followed prospectively for BMD and fractures||Prevalent fractures||Fractures|
Total hip fractures
|Study||Sample size/Follow-up||Inclusion Criteria||Exclusion Criteria||Independent assessment of risk factors & fractures||Statistical Method||Complete follow-up|
|Johnell et al., 2004 (98)|
|Population based longitudinal|
N = 1,918 men & women
Follow-up = 5 years
(osteoporotic fracture on future fractures & mortality)
|Patients in Malmo with an osteoporotic fracture @ the spine,|
shoulder, or hip identified from radiograph (1990–1994)
|Incomplete radiographic follow-up||Not stated for evaluation of incident fractures||Poisson model to calculate rate of new fractures after a fracture taking mortality into account. The rate was calculated as a function of age, sex, & time after fracture.||Lost to follow-up due to moving out of Malmo: 2.5% of men & 2.7% of women|
|Van der Klift et al., 2004 (103) Rotterdam Study||Population based longitudinal N = 1,377 men & 1,624 women Mean follow-up = 6.3 years||Age ≥ 55 years living in Ommoord, Rotterdam & Had 2nd follow-up visit in the Rotterdam study||No baseline visit; data on one or more risk factors were missing||Not stated for morphometric evaluation of baseline & follow-up radiograph of thoracolumbar spine Interviews re medical history, drug use, diet, falls & non-vertebral fractures after age 50 years||Test for significance of risk factors on incident vertebral fracture, using unadjusted & adjusted (age, BMD, prevalent nonvertebral fractures) models of logistic regression||71% of original selected subjects had complete data.|
Accounted for exclusions
|Szulc et al., 2005 (97) MINOS||Population-based longitudinal N = 759 men Mean follow-up = 7.5 years||Men age ≥ 50 years & had DXA|
absorptiometry @ baseline in MINOS study
|Had high trauma fractures; fractures of figures, toes or skull; fractures before 40,||Not stated for evaluation of incident fractures or DXA absorptiometry||Logistic regression to determine increase in fracture risk/SD decrease in BMD adjusted for risk factors. Did not use Cox proportional hazard model.||Analysis included 96% of original cohort who had reliable data about fracture incidence|
|Colon-Emeric et al., 2003 (105) US||Two cohort studies† EPSE n = 10,680|
BHS = 549 Mean follow-up EPSE 6 years BHN 1.6 years Outcome: self-reported nonhip skeletal fractures during follow-up
community dwelling adults≥65 yrs BHS included community dwelling men & women >65 yrs admitted to hospital with acute hip fracture
|Excluded: patients with no follow-up visits or who reported a history of nonhip fractures.||Not applicable||Survival analysis; Cox proportional Hazard model stratified by site.|
Model adjusted for race, sex, age, BMI, stroke, cancer, difficulty walking acrossa room etc.
|All included patients|
|Roy et al., 2003 (153)|
|Prospective Multicenter longitudinal N=3,173 men & 3,402 women Mean follow-up = 3.8 years||Age 50–79 years randomly sampled from population registers in study centres & had repeat lateral thoraco-lumbar spine radiographs – completed questionnaire||<50 years||Yes, baseline& follow-up radiographs evaluated morphometrically & qualitatively at central radiological facility||Poisson regression used to explore relationship between patient risk factor variables & incident vertebral fracture, adjusted for age & centre|
|Cauley et al., 2004 (147)|
|N = 317 of 523 men from STORM study 2,067 women from SOF study who had info on prevalent vertebral fractures||Caucasian men & women in 2 separate longitudinal cohort studies (population-based list)||Inclusion: living in community of Monongahela Valley (Pittsburgh) - walk without assistance of another person Exclusion: people with bilateral hip replacement||Morphometric definition of vertebral fractures given.|
No masking of assessment stated.
|Chi-square test for categorical variables. Logistic regression analysis to calculate OR of having a vertebral fracture per 0.01g/cm2 decrease in BMD.|
Analyzed effect of gender & age.
|Schuit et al., 2004 (145)|
|Population-based cohort study|
N = 3,075 men & 4,731 women Mean follow-up = 6.8 years (BMD & nonvertebral fractures)
|Inhabitants of Ommoord ≥55 years who consented to participate Underwent Clinical exam & DXA BMD of FN||No informed consent for follow-up registration||Reported events verified & coded independently by research physicians & confirmed by a medical expert.||RR for first fracture associated with 1 SD decrease in femoral neck BMD using a Cox proportional hazard model||Results reported entire cohort.|
FN BMD available in 74.2% of participants
|Van Potelbergh et al., 2003 (196)||Longitudinal population study N = 214 men||Healthy; age > 70 years recruited from population register of a semi rural community in Belgium||Past or current disorders or treatments that may affect androgen status &/or bone metabolism; incomplete data|
|Pande et al., 2000 (146)||Case-control study N=100 men - study subjects & 100 age-matched controls||Study subjects Consecutive men>/=50 years old admitted to a UK hospital with a low trauma hip fracture||Excluded trauma fractures & residence outside Cornwall, & active malignancy||Not stated for DXA BMD measurements @ L-spine & proximal femur; hip axis length (HAL) recorded by automated software||Association between BMD, HAL and risk of hip fracture determined using logistic regression, adjusted for age, & subsequent height & weight||62 study patients & 100 controls had data concerning BMD. Accounted for lack of data|
|Naves et al., 2003 (106) EVOS (Spain)||Cross sectional. N = 316 women and 308 men Follow-up 8 years (Effect of vertebral on risk of further osteoporotic fractures)||Random selection from register of Oviedo in Spain; Age ≥ 50 years.||Not stated||Semi quantitative & morphometric evaluation of lateral radiographs of dorsal & lumbar spine taken @ baseline & 4th year of follow-up Blinded for incidence of new fractures||RR of different osteoporotic fractures (with vs no vertebral fractures). K-M survival curves. Cox multiple regression to compare survival (with vs no prevalent vertebral fractures.||Analysis included all patients.|
|Van Staa et al., 2002 (220)|
|Population-based –register N= 119,317 women & 103,052 men from General Practice Research Database GPRD||All patients age ≥ 20 years registered in GPRD& had a fracture recorded in 1988–1998||Not stated||Not stated||Standardized incidence ratios of observed to expected number of fracture cases during follow-up Adjusted for age & incidence||Analysis included all patients|
|Pongchaiyakul et al., 2005 (99) DOES|
Follow-up median 10.2 (SD 4) years
|Randomly selected from database of a large population (Dubbo) study – residence in an isolated Australia city||Inclusion: -60 years of ageas of June 1989 -Free of anyillness likely toaffect bonemetabolism||Presence of vertebral deformity read in a masked fashion to BMD. Precision of BMD: 1.3–1.5%||Cox proportional hazards model, adjusted for age, sex, BMD & body weight||100%|
|Ismail et al., 2001 (100) EVOS (Europe)||Multicenter longitudinal N = 6,344 men & 6,788 women Median follow-up = 3 years||Age ≥ 50 years recruited from population registers in 36 European centres through stratified sampling||Not stated||Not stated for Morphological identification of spine deformity in radiographs & confirmation of reported fractures||Cox proportional hazard regression analysis to assess the predictive risk of vertebral deformity on future limb fractures,||Reported results for all patients|
|Hasserius et al., 2003 (101)|
|Longitudinal N = 298 men & 300 women Randomly selected from population of city of Malmo|
Follow-up = 10 years
(spine deformity on risk of future fractures & mortality)
|Age 50–80 years||Not stated||Not stated for morphological identification of spine deformity in radiographs|
Mortality & fracture (all types, fragility, hip) incidence
|Cox proportional hazard regression model Multivariate analysis adjusted for age, alcohol intake, smoking, general health & previous hip fracture||Reported results for all patients.|
|Leifke et al., 2005 (221)||Case-control N = 27, men and 12 controls (after non-immobilizing stroke, no fracture)||Study group:|
1–3.5 months after a minimal trauma hip fracture; ≥ 65 years
|Men with secondary osteoporosis||Laboratory assay of sex hormones (T, non-SHGB-bound T, E2, iPTH||Yes||100%|
|Mellstrom et al., 2006 (171) MrOS (Multinational)||Multicenter Cross-sectional N = 2,908 men||Men age 69–80 years randomly selected from population registers in Sweden, Hong Kong, & US; able to walk without aids,||Bilateral hip prosthesis||Not stated for: BMD (total hip, femoral trochanter & L-spine. Assay of serum total T, E2, & SHBG Self-reported fractures.||Pearson correlation for univariate associations.|
Linear regression for independent predictors.
Odds ratios to determine predictive values.
|All 2,908 patients included in analysis,|
* CV of DXA BMD tests ranged from 0.5% to 3%
|Barret-Connor et al., 2000 (170) Rancho Bernardo Study||Longitudinal study|
N = 352 men & 288 postmenopausal women
Mean follow-up = 8.4 years
|Community – dwelling, ambulatory residents of Rancho Bernardo; Caucasian; age ≥ 50 years.||Women on estrogen;||Masked assessment not stated for lateral radiograph of thoracic & L-spine. Blood assay of free & bioavailable sex hormones||Pearson’s correlation used to assess association between hormone levels & age; Mann-Whitney test for difference by fracture status using median age.||14 of original cohort excluded because of suspected hormone use, endocrine disease, or uninterpretable radiograph|
|Kanis et al., 1999(155)|
|Case control study|
N = 730 men Identified by surveillance of hospitals, clinics, & nursing homes + 1,132 age-matched controls
|Caucasian, age≥ 50 years, in catchment area in Southern Europe & had a hip fracture||Poor mental sautés, refusal, or concurrent illness||Questionnaire : assessed height, body weight, physical activity, mental score, intake of alcohol, tobacco, exposure to sun light||RR estimated from OR & adjusted using logistic regression models. The multivariate analysis used unconditional logistic regression.||All study subjects and controls included in the analysis|
|Kanis et al., 2002(222) NHANES III Sweden||Population-based study||Fracture risks obtained from Malmo and applied to the population of Sweden||Calculated 10-yr probability in 5-year age groups using hazard of first fracture @ each of the hip, spine, shoulder & distal forearm & the death hazard|
|Meyer et al., 1998 (156) (Norway) Hip fracture||Longitudinal study|
N = 19,151 men & 19,938 women Mean follow-up =11.6 years
|Men & women born in 1925–1940, living in 3 Norwegian counties & attended 3 consecutive health examinations by the National Health Screening Service||Not stated||Hip fractures (cervical & trochanteric) identified from registers& verified through medical records||Age adjusted incidence rates using Cox proportional hazards regression; RR from multivariate analysis; association between weight variability & trend in weight change by Pearson’s correlation.||All patients included in analysis.|
|Langlois et al., 1998 (157)|
|Longitudinal study N = 2,413 men|
Follow-up = 8 years
|Caucasian community-dwelling men in 3 US counties; age ≥ 67 years; & participated in the baseline interview.||Non-white men; no match to Medicare hospitalization file, had previous hip fractures, or missing body weight or height data.||Identified men hospitalized for hip fracture during follow-up.|
BMI & weight change determined.
|Adjusted RR for hip fracture by category of weight change from age 50 years - from a Cox proportional hazard model stratified by community||Results reported for all 2,413|
|Gonnelli et al., 2005 (143)||Cross sectional study|
N = 401 mens consecutively referred
|Referred for assessment of bone status Jan 2000–Dec2002||Excluded : extensive exclusion criteria including >15% below or 30% above ideal body weight, cannot have spine DXA BMD test, secondary causes of osteoporosis, etc||Not stated for|
(DXA BMD @ spine, FN, total hip, trochanter & intertrochanter & ultrasound).
|Pearson correlation coefficient (BMD & fracture risk), Receiver Operator characteristic analysis & AUC||Results reported for all 401 patients|
|Kudlacek et al., 2000 (148)|
|Cross sectional N = 136 men & 337 women||Patients referred to a single center for screening or diagnosis of suspected osteoporosis||Secondary osteoporosis or receiving medical treatment for osteoporosis, history of severe trauma||Not stated for fracture assessment or BMD measurements|
Lumbar spine BMD by QCT
|Logistic regression model to estimate probability of a fracture depending on BMD, matched & compared for sex groups.||Reported for all patients|
|Kelsey et al., 2005 (31)|
|Case control N = 192 consecutive men and women Controls : 2,402 randomly selected from KP medical centre from same period (not matched)||Pelvis fracture identified from radiology or medical reports, confirmed by radiograph, bone scan or MRI. Controls-||Prior pelvis fracture after age 45 ; Fractures from disease such as Paget’s or cancer.||Not stated|
Risk factors by standardized questionnaire by trained interviews. Measurement of physical functioning
|Odds rations adjusted using unconditional logistic regression for sampling variables of age group, sex, & race/ethnicity||Reported for all patients|
|Holmberg et al., 2005 (154)|
|Longitudinal N = 22,444 men (Mean age 44 years) & 10,902 women (mean age 50 years) Follow-up = 16 years for men & 11 years for women||Population of Malmo, Sweden between 1974–1984 who consented to the study||Identification of hip fracture from radiology register (excluded high energy fractures)|
Questionnaire on risk factors (fracture history, family history, lifestyle factors)
|Age adjusted Cox proportional hazard model to analyze each variable and in categories. Analysis in a final multivariate Cox regression model|
|Van Staa et al., 2000 (161)|
|Retrospective cohort study N = 244,235 (101,599 men) study patients & 244,235 matched controls from the GPRD||Study patients: Received ≥ 1 Prescription for oral corticosteroid|
Grouped according to dosage of glucocorticoid
|Not stated||Not stated||Cox proportional hazards models to calculate adjusted RRs in comparison between oral corticosteroid dose groups. Poisson regression was used in the analysis of the cumulative vs daily doses||All patients included in analysis|
(outcomes: risk of vertebral, non-vertebral, & hip fractures)
|Van Staa et al., 2001 (166)|
|Retrospective cohort study N = 170,818 (77,763 men) inhaled corticosteroid users, 108,786 bronchodilator users, 170,818 controls from the GPRD Follow-up: from 91 days after last inhaled prescription until a fracture or censored||Age≥18 years Inhaled corticosteroid or bronchodilator group: received ≥ 1 prescription for the respective drug during the study period||Inhaled corticosteroid user who received a prescription for oral corticosteroid 6 months prior or 91 days after the last inhaled corticosteroid prescription.||Not stated||Adjusted RRs for fractures estimated using Cox proportional hazards models that included age, gender, & selected confounding variables.||All patients included in analysis|
(Outcome: risk of vertebral, nonvertebral, hip, & forearm fracture)
|Hubbard et al., 2002 (167)||Case-control study N = 16,341 (3,432 men) study patients& 29,889 matched controls from GPRD||Study group: All patients in GPRD with recorded diagnosis of hip fracture.|
Controls: 2:1 ratio matched for age, sex, general practice, & start date for collection of prescribing data.
|Not stated||Not stated||Relationship (ORs) between inhaled corticosteroid & hip fracture quantified by conditional logistic regression. Bivariate & multivariate models used to determine impact of confounders.||Not stated|
Outcome: relationship between exposure to inhaledcorticosteroid & risk of hip fracture)
|Hubbard et al., 2006 (164)|
|Prospective cohort analysis N = 1,671 patients (868 men) from computerized general practice records of 31 practices (January 1995 – February 1999)||Age≥ 75 years; Diagnosis of asthma, COPD, or both||Not stated||Not stated||Cox regression modeling for fracture risk of people with vs those without exposure; controlled for age & gender. Multivariate model to explore potential confounders||Included all patients in analysis. (Outcomes: risk of all fractures (exposed vs non-exposed) (oral, inhaled, injected)|
|Vestergaard et al., 2003 (160)|
N = 6,660 study patients (1,974 men) + 33,272 age-matched population controls (no fracture)
|Study patients: In County Hospital Discharge Register ; a 1st diagnosis of hip fracture (1994–2001)||Had hip fracture before 1993; had an address outside the county; residence in county <5 years.||Not stated||OR for hip fractures based on cumulative dose by conditional logistic regression. Univariate & multivariate analysis for potential confounders||Included all patients in analysis|
|Vestergaard et al., 2005 (163)|
N = 124, 655 study patients (60,084 men) +373,962 controls matched on age & gender (random selection)
|Study patients: All people in Denmark who had a fracture in year 2000 (National Hospital Discharge Register||None stated||Not stated||Conditional logistic regression for crude & adjusted ORs of any fractures (adjusted for exposure variables, comorbid conditions & other drug use.||Included all patients in analysis|
(oral, inhaled, injected, or topical corticosteroid)
|Steinbuch et al., 2004 (162)|
|Retrospective cohort study N = 17,957 (7,509 men) + 17,957 controls matched for age, sex, & date of claim from same database||Study patients: in an admin claims database, age 18–64 years, had ≥ 1 claim for oral glucocorticoid 1995–1996; continued enrolment in the drug & medical plan during follow-up||Study patients: excluded patients with a pharmacy claim for an injectable steroid during the study period||Not stated||RR for of fracture (study group vs controls) estimated using a Cox proportional hazards model, adjusted for age group, sex, prior fracture, & prior exposure; & for amount, duration & pattern||All patients included in analysis (Outcomes: risk of hip, non-vertebral, vertebral, & forearm fracture during 1 year after first exposure)|
|Donnan et al., 2005 (165)|
|Retrospective cohort study N = 20,226 (men 40.5% of total person years) + 248,723 controls January 1993–January 1997||Age> 18 years from an administrative claims database; ≥ 1 dispensed prescription for oral corticosteroids; Controls: unexposed to glucocorticoids||None stated||Not stated||Regression analysis for RR of fractures (vertebral, non-vertebral, hip & wrist) per 1,000 person years for exposed vs non-exposed to glucocorticoids. Population attributable risk estimated from adjusted RR||All patients included in analysis|
(Outcomes: risk of non-traumatic vertebral & non-vertebral fracture during follow-up
|Study||Design||Age||BMD||2 BMI Kg/m2 or Weight loss||Previous fractures/V. deformities||Sex Hormones||Other Risk Factors|
|Klotzbuecher et al., 2000 (94)||Meta-analysis 33 studies||Prior fracture @ any site predicts future fracture RR 1.9–2.6 (men & women)|
|Kanis et al., 2004 US (95)||Meta-analysis (RR similar for men & women)||Low BMD explained a minority of risk for any fracture (8%), 22% hip fracture.||RR 1.86, 95% CI, 1.75–1.98 for hip & osteoporotic fractures independent of BMD|
|Haentjens et al., 2003 (96)||Meta-analysis 9 cohort studies (Colles or spine)||Colles RR 3.26|
Spine RR 3.54
|De Laet et al., 2005 (152)||Meta-analysis|
|Kanis et al., 2004 (158)||Meta-analysis||Parental history of fracture increased risk of fracture|
|Johnell et al., 2004 (98)||Longitudinal N=268 men & women||Osteop. Fracture RR 3.7–125|
|Van der Klift et al., 2004 (103) Rotterdam Study (Incident vertebral fractures)||Prospective population-based single-cohort|
N =1,377 men & 1,624 women Follow-up 6.3 years
|Women: age RR (1.8-2.6) depending on age||Men Low L-spine BMD RR 2.3 (1.6–3.3)|
Women low L-spine BMD RR 2.1 (95% CI, 1.6–2.6)
|Men-prevalent vertebral (RR 2.2)& non-vertebral fracture (RR 2.4)||Women: menopause before at or before age 45 years||Women use of walking aid (RR 2.5)|
Current smoking (RR 2.1)
|Szulc et al., 2005 (97) MINOS||Prospective cohortn=759 >50 years Follow-up, 7.5yrs||OR 1.29 (1–1.64, P =.05) per 10 year increase in age Risk @ age 75 yrs 3x risk @<55 yrs||Adjusted OR 1.28–1.89 per SD decrease in BMD AUC 0.643–0.712||Unadjusted fracture risk with in body weight OR 1.15 per 5 kg weight (1.03–1.28) P<.02)||2-fold in riskregardless of site of previous fracture|
|Colon-Emericet al., 2003 (105) US BHS 1.6 yrs, EPESE 6 yrs||EPSE † n = 10,680 BHS n = 549 Mean age 73.8 yrs & 80.9 years respectively||Hip fracture linked to 2.5-X increase in nonhip fractures in men & women|
|Naves et al., 2003 (106)|
|Longitudinal N = 308 men &316 women≥ 50 years old Men & women.|
Follow-up of 8years
|Prevalent V fracture predict vertebral (RR 4.7), hip (RR 6.7) & Colles fracture (RR 3)|
|Van Staa et al., 2002 (220)|
|Longitudinal N = 103,052 & men, 119,317 women||Any prevalent fracture SIR 4.2–13.4|
|Pongchaiyakul et al., 2005 (99)|
N=114 men and 186 women
Follow-up men 10.5 yrs, women 10 yrs
|Asymptomatic deformity associated with higher incidence of subsequent fractures & mortality in both sexes|
|Ismail et al., 2001 (100)|
|Multicenter Longitudinal N = 6,344 men & 6,788 women age ≥ 50 yrs Median follow up = 3 years||Vertebral deformity is strong predictor of hip & limb fracture in women but not in men|
|Hasserius et al., 2003 (101)|
|Longitudinal N = 298 men & 300 women age ≥ 50 yrs Follow-up = 10 years||V. deformity predicted mortality & risk of any fracture in women & men|
|Roy et al., 2003 (153)|
|Longitudinal N=3,173 men Mean follow-up 3.8 years (risk of vertebral fracture)||Risk increased with age||Not studied||Men in lowest quintile wt compared with others, RR for V. fracture 1.99 (95% FI 1.01–3.93)||Not studied||Smoking & alcohol or milk intake not associated with risk ofvertebral fractures|
|Cauley et al., 2004 (147)|
|(n=317 Caucasian men Age & 2,067 Caucasian women >50 yrs old||Prevalence of vertebral fractures increased with age in both men & women||0.1g/cm2 in BMD associated with 30–40% in men & 60–70% in women of risk of V. fracture|
|Kanis et al., 1999 (155)|
|Case control N=730 men ≥50 yrs w a hip fracture & 1,132 controls Interview: re Risk factors for hip fracture in men.||6.8% in fracture risk with each unit in BMI.|
RR = 0.68 (P< .009)
|Gender at each threshold probabilities higher in women than in men, esp. hip & spine fractures||Recreational Physical activity RR = 0.64 (P<.002)|
|Kanis et al., 2002 (222) NHANES III|
|10-yr probabilities of fractures by age & T-score Threshold for T-score 0.577g/cm2 osteoporosis 0.740g/cm2 low bone mass||10-year probabilities for fractures (except forearm) increased with age, effect independent of BMD.||10-year probabilities of fractures decreased with T-score|
|Meyer et al., 1998 (156) (Norway)|
|Longitudinal study N = 19,151 men & 19,938 women|
Mean age 49 yrs
Follow-up: 12 year
|Men & women who had the most weight variability had increased risk of hip fractures independent of BMI.|
|Langlois et al., 1998 (157)|
|Longitudinal study N = 2,413 white men in community Age ≥ 67 years||>10% wt loss since age 50 associated with 2-fold increase in risk of hip fracture|
|Gonnelli et al., 2005 (143)||X-section n=401 men||Hip BMD predicts hip fracture OR 3.42|
|Schuit et al., 2004 (145)|
|T-score < –2.5 identified only 21% of nonvertebral fractures in elderly men & 44% in elderly women.|
|Van Potelbergh et al., 2003 (196)||(n=540 men, 2,264 women) 65-89 yrs for men||x|
|Pande et al., 2000(146)||Cross sectional study|
N = 62 men with low trauma hip fracture > age 50 years 100 controls
|Adjusted for age, height & weight, OR for fracture per 1SD reduction in BMD =1.8 L-spine, 3.1 for femoral neck, 3.9 for trochanter, 4.0 for intertrochanter area 3.7 for wards triangle||No association between hip axis length & risk of fracture.|
|Kanis et al., 2005(173)||Meta-analysis 10 studies N =15,400 men & 43,832 women 250,000 person-years||Current or history of smoking risk of fracture >explain edby|
|Kanis et al., 2005 (172)||Meta-analysis 3 studies N = 5,939 men & 11,032 women||>2 units alcohol/d ay risk of any, osteop, & hip fractures (men & women)|
|Leifke et al., 2005(221)||Systematic review + cross sectional study N= 27 after hip fracture Control = 138 healthy youth & 110 (60–80 yr old males)||Review = conflicting finding X-sectional study: hip fracture associated with >2 SD below control|
|Mellstrom et al., 2006 (171) MrOS (Sweden)||Cross-sectional N=2,908 men Mean age 75.4 yrs DXA BMD, fractures rates & serum sex steroids||Free testosterone predicts BMD & osteoporotic fractures & vertebral fractures|
|Barret-Connor et al., 2000 (170) Rancho Bernardo Study||Longitudinal N = 352 men, 288 women, ≥age 50 yrs.|
Predictors of vertebral fractures
|Not studied||Low total &bio available estradiol associated with 4–5 fold odds ofvertebral||No|
|Kanis et al., 2005 (223)|
151,957 person years
|Meta-analysis of 6 prospective cohort studies including CaMos N = 39,563 men & women (69% women)||Low intake of milk was it was associated with increased risk of osteoporotic fracture only after age 80 yrs.|
|Kudlacek et al., 2000(148)||Cross sectional study|
N = 136 men & 337 women mean age 60.7 & 59.7 yrs respectively 52 & 96 had a spine fracture
|Men fracture @ a higher BMD level than women OR for gender 3.1|
|Kelsey et al., 2005 (31)|
|Case control N = 192|
consecutive men & women with pelvic fracture, age =/>45 years 2,402 controls @ 5 Kaiser
Permanente centres 21% of cases & 50% controls from racial/ethnic minority groups
|Propensity to fall & indications of frailty associated with increased risk.|
Need help to perform physical function.
Inactivity during leisure time in past year. Self reported diseases:
|High BMI OR =|
0.65 (0.52–0.81) per 5 units increase - protects loss of bone mass
|Number of fractures since age 45 Adjusted OR = 1.42 (1.03–1.96)|
Maternal history of hip fracture OR = 1.72 (1.02–2.90)
|Recent use of menopausal hormone OR = 0.55 (0.33–0.91) Protectsloss of bonemass Hysterectomy OR = 1.75 (1.15–2.66)||Current smoker OR = 2.17 (1.34–3.52)|
|Holmberg et al., 2005 (154)|
|Prospective population-based observational study N=22,444 men & 10,902 women mean age 44 (27–61) Followed-up: 16 yrs for men & 11 yrs for women Cox regression model||Similar risk factors for cervical hip fractures in men except admission for mental disorder is not a risk factor for cervical fracture||For hip fracture in men: High BMI protective in RR 0.63 (0.53–0.76, P=.0001) Sleep|
disturbance RR = 1.84 (1.25–2.70, P = .002) Multivariate analysis: All 7 variables were associated with risk of hip fractures & cervical hip fractures Diabetes strongest association to cervical hip fracture.
|Previous fracture increased risk in women RR 4.76 (2.74–8.26) for hip fracture||Self-reported diabetes strongest predictor RR 4.07 RR 7.75 (4.37–13.7, P = .001) High BP & elevated resting pulse. Elevated serum glutamy transferace RR = 1.45 (1.28–1.65)||Hip fracture in men|
Current smoking or self rated poor health RR 2.72 (1.94–3.80, P=.001) Hospital admission for mental disorder RR = 2.64(1.46-4.76, P=.01)
|Vestergaard et al., 2003 (160)||Prospective case-controlled|
N = 6,660 people with hip fractures & 33,272 matched controls 5 year study
|Use of oral glucocorticoid – 7.5mg/day × 6 months OR 1.36 (95% CI 1.19–1.56)|
>1500mg: OR 1.65 (95% CI 1.43–1.92)
|Vestergaard et al., 2005 (163)|
|Prospective community-based case-controlled N=124,655 & 373,962 matched controls||Oral corticosteroid>2.5 mg/day increased risk of hip, spine & forearm, but no increase for inhaled corticosteroid|
|Steinbuch et al., 2004 (162)|
|Retrospe ctivecohort study N = 17,957 (7,509 men) + 17,957 controls Patients had a claim or oral corticosteroid||Oral corticosteroid increased risk of vertebral, nonvertebral & hip fractures RR ranged from 1.53 to 3.41 depending on dose & duration|
|Donnan et al., 2005 (165)|
|Retrospective cohort study population ≥18||Inhaled corticosteroid risk of all fractures RR 1.9 (1.68–2.16)|
|Van Staa et al., 2000 (161)|
|Retrospe ctivecohort study N = 244,235 oral corticosteroid users & 24,235 controls||Oral corticosteroid – effect on risk of fractures was dose dependent; risk doubled as dose from 2.5 mg to 7.5 mg|
|Van Staa et al., 2001 (166)|
|Retrospective cohort study N = 17,818 inhaled corticosteroid users (55% men), 108,786 bronchodilator users, and 170,818 controls||Inhaled corticosteroid increased risk for vertebral, non vertebral & hip fractures RR ranged from 1.15 to 1.51|
|Hubbard et al., 2002 (167)||Case control study N = 16,341 people with hip fractures (21% men) (meanage 79 years) & 29,889 matched controls||Inhaled corticosteroid increased risk forhip fractures OR 1.26 (95% CI1.17–1.36)|
|Hubbard et al., 2006(164)||Case control study N = 1,671 people with asthma or COPD (mean age 80.6 yrs)||Inhaled corticosteroid increased risk for any fractures HR 2.53 (95% CI 1.65–3.89)|
This report should be cited as follows:
Medical Advisory Secretariat. Utilization of DXA bone mineral densitometry in Ontario: an evidence-based analysis. Ontario Heath Technology Assessment Series 2006; 620)
All inquiries regarding permission to reproduce any content in the Ontario Health Technology Assessment Series should be directed to MASinfo/at/moh.gov.on.ca
All reports in the Ontario Health Technology Assessment Series are freely available in PDF format at the following URL: www.health.gov.on.ca/ohtas
Print copies can be obtained by contacting MASinfo/at/moh.gov.on.ca
All analyses in the Ontario Health Technology Assessment Series are impartial and subject to a systematic evidence-based assessment process. There are no competing interests or conflicts of interest to declare.
All Medical Advisory Secretariat analyses are subject to external expert peer review. Additionally, the public consultation process is also available to individuals wishing to comment on an analysis prior to finalization. For more information, please visit http://www.health.gov.on.ca/english/providers/program/ohtac/public_engage_overview.html
The Medical Advisory Secretariat
Ministry of Health and Long-Term Care
20 Dundas Street West, 10th floor
The Medical Advisory Secretariat is part of the Ontario Ministry of Health and Long-Term Care. The mandate of the Medical Advisory Secretariat is to provide evidence-based policy advice on the coordinated uptake of health services and new health technologies in Ontario to the Ministry of Health and Long-Term Care and to the healthcare system. The aim is to ensure that residents of Ontario have access to the best available new health technologies that will improve patient outcomes.
The Medical Advisory Secretariat also provides a secretariat function and evidence-based health technology policy analysis for review by the Ontario Health Technology Advisory Committee (OHTAC).
The Medical Advisory Secretariat conducts systematic reviews of scientific evidence and consultations with experts in the health care services community to produce the Ontario Health Technology Assessment Series.
To conduct its comprehensive analyses, the Medical Advisory Secretariat systematically reviews available scientific literature, collaborates with partners across relevant government branches, and consults with clinical and other external experts and manufacturers, and solicits any necessary advice to gather information. The Medical Advisory Secretariat makes every effort to ensure that all relevant research, nationally and internationally, is included in the systematic literature reviews conducted.
The information gathered is the foundation of the evidence to determine if a technology is effective and safe for use in a particular clinical population or setting. Information is collected to understand how a new technology fits within current practice and treatment alternatives. Details of the technology’s diffusion into current practice and information from practicing medical experts and industry, adds important information to the review of the provision and delivery of the health technology in Ontario. Information concerning the health benefits; economic and human resources; and ethical, regulatory, social and legal issues relating to the technology assist policy makers to make timely and relevant decisions to maximize patient outcomes.
If you are aware of any current additional evidence to inform an existing Evidence-Based Analysis, please contact the Medical Advisory Secretariat: MASInfo/at/moh.gov.on.ca.. The public consultation process is also available to individuals wishing to comment on an analysis prior to publication. For more information, please visit http://www.health.gov.on.ca/english/providers/program/ohtac/public_engage_overview.html
This evidence-based analysis was prepared by the Medical Advisory Secretariat, Ontario Ministry of Health and Long-Term Care, for the Ontario Health Technology Advisory Committee and developed from analysis, interpretation and comparison of scientific research and/or technology assessments conducted by other organizations. It also incorporates, when available, Ontario data, and information provided by experts and applicants to the Medical Advisory Secretariat to inform the analysis. While every effort has been made to do so, this document may not fully reflect all scientific research available. Additionally, other relevant scientific findings may have been reported since completion of the review. This evidence-based analysis is current to the date of publication. This analysis may be superceded by an updated publication on the same topic. Please check the Medical Advisory Secretariat Website foralist of all evidence-based analyses: http://www.health.gov.on.ca/ohtas
1Odds ratios are reliable estimates of the relative risk when they include unbiased samples of cases and controls (i.e., population-based studies) and when they are used in studies of rare events (i.e., the end point occurs in <15% of the subjects) (Haentjens 2003)