When to use FRAX and DXA--and what they mean.
How FRAX and DXA work
World Health Organization (WHO) experts devised a simple online algorithm that lets clinicians plug in easily obtained clinical variables (see Note 4), click a button, and get a 10-year prediction of (1) hip fracture risk and (2) major osteoporotic fracture risk (spine, hip, wrist, upper arm). Femoral neck bone mineral density (BMD) can be added to the algorithm, but the calculation works without it. In people with HIV, FRAX does not predict DXA-determined osteoporosis or prevalent fracture. (3)
DXA--short for dual-energy x-ray absorptiometry--beams low-dose x-rays with two distinct energy peaks through the target bone. (5) Soft tissue absorbs one peak while bone absorbs the other. Subtracting the soft-tissue value from the total value yields BMD. A DXA T score compares a patient's BMD with that of a young adult of the same gender, as follows:
* A value above -1 means normal BMD
* A value between -1 and -2.5 means osteopenia
* A value below -2.5 means osteoporosis
A DXA Z score indicates a person's BMD compared with people the same age, the same size, and the same gender. Z scores at or below -2 indicate low BMD for age, but Z scores are not used to diagnose osteoporosis. (1)
When and how to use FRAX and DXA
New bone screening guidelines for adults with HIV start by dividing people into those with no fragility-fracture risk factors and those with risk factors, which include (1) history of fragility fracture, (2) glucocorticoid treatment for more than 3 months, and (3) a high risk of falls (Figure 1). (1) People with fragility-fracture risk factors should get a DXA scan; people without risk factors should be considered for FRAX or DXA according to age or menopause status (Figure 1). (Fragility fractures are those following minimal trauma and usually affecting the hip, spine, or wrist. (6))
What do FRAX and DXA results say about planning care and planning follow-up? A FRAX score predicting a 10% or lower fragility-fracture risk in 10 years means patients should be counseled on calcium and vitamin D intake and on lifestyle issues that may affect bone changes--like alcohol, smoking, and weight-bearing exercise (Figure 2). (See the article starting on page 20.) People with a FRAX result above 10% should have DXA to measure BMD, and those with a FRAX at or above 20% should undergo DXA and may be candidates for bisphosphonate therapy after the clinician excludes secondary causes of osteoporosis.
For people without a prior fracture and a FRAX below 20%, those with a lowest DXA-calculated T score above -2.5 should get calcium, vitamin D, and lifestyle advice (Figure 2). In the United States (1) a T score at or below -2.5 at the femoral neck, total hip, or lumbar spine, or (2) a T score between -1.0 and -2.5 and a FRAX score at or above 20% or at or above 3% at the hip, or (3) a hip or vertebral fracture should lead clinicians to consider bisphosphonates--plus vitamin D, calcium, and lifestyle advice--after excluding secondary causes of osteoporosis or low BMD.
HIV-positive people with a FRAX score at or below 10% should have a follow-up FRAX in 2 to 3 years or if a new fracture risk factor develops (Figure 3). People with a FRAX below 20% and no fracture history should have another DXA in 1 to 2 years if they have advanced osteopenia (T score -2.00 to -2.49) and in 5 years if they have mild to moderate osteopenia (T score -1.01 to -1.99). People who start bisphosphonates need a follow-up DXA after 2 years.
Aiming to figure how often DXA scans should be repeated, Spanish researchers analyzed 1639 DXAs in 391 HIV-positive people to gauge abnormal BMD progression risk. (7) Everyone in the analysis had at least two DXAs from 2000 through 2009. Initial scans indicated that 29% had normal BMD, 50% had osteopenia, and 22% had osteoporosis. Using these baseline BMD findings, the investigators divided people into tertiles at low risk, middle risk, and high risk of progression from normal BMD to osteopenia and from osteopenia to osteoporosis (Table 1). The new HIV bone guidelines (1) suggest repeat DXA intervals prudently within the progression bounds found in the Spanish study.
FRAX underestimates fracture risk in older HIV+ men
Shortly after release of the new HIV bone guidelines, (1) researchers reported results of a large comparison of 50-and-older men with or without HIV indicating that FRAX underestimates fracture risk more in men with HIV "and does not discriminate well between those at risk and not at risk for future fracture." (3) The new guidelines reflect this result in recommending DXA--not FRAX--for men over 49, (1) but the FRAX study raises concern because no one assumes a clean cutoff at age 50 that renders FRAX inaccurate over 50 and accurate under 50.
With a team of US colleagues, Columbia University's Michael Yin (a member of the guideline panel (1)) studied 50- to 70-year-old HIV-positive and negative men in the Veterans Aging Cohort Study Virtual Cohort (VACS-VC) who had complete data from 2000 for all FRAX algorithm variables except secondary osteoporosis and parental hip fracture. For 7064 men with HIV and 17,387 without HIV, the investigators searched the VACS database to determine how many men had a hip, spine, wrist, or upper arm fracture from 2001 through 2010.
The HIV-positive and negative groups both averaged 56 years in age, and similar proportions were black (48% and 46%) or white (44% versus 45%). Veterans with HIV weighed significantly less than veterans without HIV (average 79 versus 89 kg, P < 0.01), and higher proportions of positive men had a previous fracture (2.4% versus 1.8%, P = 0.003) and used glucocorticoids (0.3% versus 0.1%, P = 0.02). But lower proportions of HIV-positive vets smoked (5.2% versus 6.3%, P < 0.01) and had rheumatoid arthritis (2.7% versus 3.5%, P < 0.01). (These are all variables in the FRAX equation.)
Osteoporotic fracture incidence in the 10-year study period proved significantly higher in veterans with than without HIV (4.61% versus 3.50%, P < 0.0001), as did hip fracture incidence (1.32% versus 0.85%, P = 0.0008). FRAX-estimated 10-year fracture rates fell well short of actual fracture rates for both major osteoporotic fractures (2.85% with HIV versus 2.71% without HIV, P < 0.0001) and hip fractures (0.29% with HIV versus 0.24% without HIV, P < 0.0001). In the interview starting on page 14, Yin characterizes these findings as "illustrative but not definitive" because the analysis had to leave out two FRAX variables.
Next Yin and colleagues refigured FRAX predictions for men with HIV after checking the secondary osteoporosis box in the FRAX form as a surrogate for HIV infection. The 10-year prediction for major osteoporotic fractures rose 31% as a result, and the 10-year prediction for hip fractures jumped 67%. But both resulting predictions still substantially lagged observed fracture rates in HIV-positive vets.
The investigators calculated FRAX predictive accuracy as the agreement between observed fractures and FRAX-estimated fractures by an observed/ estimated (O/E) ratio that indicates perfect accuracy if O/E = 1.0. FRAX accuracy significantly missed the ideal ratio in HIV-negative veterans (O/E ratio 1.29, 95% confidence interval [CI] 1.19 to 1.40) and missed the mark by an even wider margin in men with HIV (O/E ratio 1.62, 95% CI 1.45 to 1.81). Predictive accuracy in men with HIV improved to a level similar to HIV-negative men when the analysis included HIV as a cause of secondary osteoporosis, but even after this correction accuracy lay significantly beyond the ideal mark (O/E ratio 1.20, 95% CI 1.08 to 1.34)
Yin and coworkers concluded that FRAX as they calculated it (leaving out two algorithm variables) has limited accuracy in HIV-positive men over 50. Checking the secondary osteoporosis box as a surrogate for HIV infection improved accuracy but still underestimated the actual fracture rate. They proposed that "the exact role of FRAX for risk stratification in HIV-infected individuals for DXA screening or pharmacologic treatment requires additional study." (3) In the interview starting on page 14, Yin suggests "there's room for a better prediction algorithm," but designing a better algorithm that uses only clinical variables won't be easy.
One way to improve FRAX accuracy in people with HIV may be combining it with the Aging Males Symptoms (AMS) scale, an 18-question form developed by the Berlin Center for Epidemiology and Health Research and available online. (8) But the findings come from a small cross-sectional study of only 50 antiretroviral-treated men with HIV and 27 controls. (9)
The AMS scale tallies symptoms related to hypogonadism (low testosterone), a recognized secondary cause of osteoporosis. A score at or above 27 is abnormal. All study participants had a DXA scan, spinal x-ray, and hormonal evaluation, and all completed the AMS questionnaire. DXA diagnosed osteoporosis in 24% of men with HIV and 4% of controls (P = 0.05); 18% of the HIV group and no controls had radiologic fractures. Calculated free testosterone levels indicated hypogonadism in 26% of men with HIV and 4% of controls (P = 0.04). Almost two thirds of men with HIV (62%) had an abnormal AMS score, compared with 41% of controls (P = 0.04).
ROC curve analysis indicated that FRAX had only 23% sensitivity but 100% specificity in identifying osteoporosis. (Sensitivity is the ability of a test to identify a condition in people with that condition; specificity is the ability of a test to identify people without a condition as not having that condition.) An abnormal AMS score yielded a sensitivity of 82.6% and a specificity of 42.9%. Combining FRAX with AMS resulted in 77.3% sensitivity and 69% specificity.
The new HIV bone guidelines (1) do not weigh in on the potential value of the AMS score in predicting bone fragility. The guidelines do recommend assessing hypogonadism as a possible cause of osteoporosis in HIV-positive men (evaluated by morning measurement of free testosterone) and women (evaluated by menstrual history, estradiol, follicle-stimulating hormone, prolactin). The guidelines list nine other potential secondary causes of osteoporosis that should be evaluated in people with HIV: vitamin D deficiency, hyperparathyroidism, subclinical hyperthyroidism, Cushing syndrome, phosphate wasting, idiopathic hypercalcuria, celiac sprue, multiple myeloma, and mastocytosis (guidelines Table 5 (1)).
If FRAX has screening limitations, why not just send everyone for a DXA scan straightaway? That's what Italian clinicians asked in a study of 163 middle-aged HIV-positive people at their center. (10) This analysis of universal DXA produced evidence that almost one fifth of HIV-positive adults with low bone density would be missed if screened by risk factor-based guidelines.
The study involved all adults seen from April 2009 through March 2011 who agreed to DXA screening. (10) Most study participants (71%) were men, and age averaged 44.2 (+/-10), so a portion of these people--those younger than 40--would not be considered for DXA or FRAX by the new bone guidelines. (1) Body mass index averaged a healthy 23.7 kg/[m.sup.2], but half of the group smoked, vitamin D levels averaged a deficient 15.2 ng/mL, and only 37% of the group exercised regularly. Defining osteopenia as a femoral or lumbar spine Z score at or below -2, the researchers diagnosed osteopenia in 32 people (19.6%). Six of those 32 (18.8%) would not be offered DXA by current Italian HIV guidelines, which call for DXA scans in anyone with HIV plus two more osteoporosis risk factors. Results were similar when the investigators used EACS guidelines. (2)
The drawbacks of a DXA-for-all strategy are clear: time and money. The procedure itself takes 10 to 30 minutes, (5) not counting time for scheduling and preparing a person for scanning. In the United States DXA scanning could cost the patient $150 to $250, (11,12) and staff time must be added to that. Staff time is the only cost for completing the FRAX algorithm. And as Michael Yin points out in the interview starting on page 14, both DXA and FRAX are less useful in predicting fracture in younger people because overall fracture risk is lower in younger people.
Screening for subclinical vertebral fracture
Clinicians may be unaware of another strong fracture predictor in people with HIV, according to the new bone guidelines. (1) One quarter of HIV-positive people break their back and never notice, and these subclinical vertebral fractures boost the risk of later vertebral fracture 5-fold and the risk of other fractures 2- to 3-fold. (13) US osteoporosis guidelines add that these hidden fractures are linked to pain, disability, deformity, and death. (13) To uncover subclinical broken vertebrae, providers should measure the height of every HIV patient 50 or older every 1 to 2 years. (1) Losing height suggests covert cracked vertebrae.
HIV bone guidelines (1) suggest lateral x-rays of the lumbar and thoracic spine or DXA-based vertebral fracture assessment to identify subclinical vertebral breaks in three groups:
* Postmenopausal 50- to 64-year old women and 50- to 69-year-old men with fragility risk factors (historical height loss [greater than or equal to] 4 cm [1.5 inches], prospective height loss [greater than or equal to] 2 cm [0.8 inches]), or recent or ongoing long-term glucocorticoid treatment)
* Women 65 to 69 years old and men 70 to 79 years old with BMD T score -1.5 or lower
* Women 70 or older and men 80 or older with BMD T score below -1.0 at the spine, total hip, or femoral neck
The guidelines go on offer advice on care for HIV-positive people at risk for fragility fracture, including antiretroviral management, bisphosphonate use, and vitamin D and calcium supplementation. (1) The article starting on page 39 of this issue considers these therapies.
(1.) Brown TT, Hoy J, Borderi M, et al. Recommendations for evaluation and management of bone disease in HIV. Clin Infect Dis. 2015;60:1242-1251. http://www.natap.org/2014/HIV/012315_01.htm
(2.) EACS European AIDS Clinical Society. Guidelines. Version 7.1. November 2014. www.eacsociety.org/files/guidelines-7.1-english.pdf
(3.) Yin MT, Skanderson M, Shiau S, et al. Fracture prediction with modified FRAX in older HIV+ and HIV- men. CROI 2015. February 23-26, 2015. Seattle, Washington. Abstract 141. http://www.croiwebcasts.org/console/player/22266?mediaType=slide Video&
(4.) The FRAX calculator is online at www.shef.ac.uk/FRAX (click Calculation Tool tab, click appropriate region, then click country). Variables that should be indicated are age, sex, race, geographic region, weight, height, previous fracture, parental fractured hip, current smoking, glucocorticoid use, rheumatoid arthritis, secondary cause of osteoporosis, and alcohol use.
(5.) American College of Radiology. RadiologyInfo.org. Bone densitometry. http://www.radiologyinfo.org/en/info.cfm?pg=dexa
(6.) American Academy of Orthopaedic Surgeons. Position statement. Recommendations for enhancing the care of patients with fragility fractures. Revised December 2009. http://www.aaos.org/about/papers/position/1159.asp
(7.) Negredo E, Bonjoch A, Gomez-Mateu M, et al. Time of progression to osteopenia/osteoporosis in chronically HIV-infected patients: screening DXA scan. PLoS One. 2012;7:e46031.
(8.) AMS--Aging Males Symptoms scale. developed by the Berlin Center for Epidemiology and Health Research. http://www.aging-males-symp toms-scale.info/
(9.) Pepe J, Isidori AM, Falciano M, et al. The combination of FRAX and Ageing Male Symptoms scale better identifies treated HIV males at risk for major fracture. Clin Endocrinol (Oxf). 2012;77:672-678.
(10.) Mazzotta, E, Ursini, T, Agostinone A, et al. Prevalence and predictors of low bone mineral density and fragility fractures among HIV infected patients at one Italian center after universal DXA screening: sensitivity and specificity of current guidelines on bone mineral density management. AIDS Patient Care STDs. 2015;29:169-180.
(11.) WebMD. Bone density: a clue to your future. http://www.webmd.com/osteoporosis/features/bone-density-clue-to-your-future?page=2
(12.) MedicineNet.com. Bone density scan. What is the cost of DXA? http://www.medicinenet.com/bone_density_scan/page6.htm#what_is_the_cost_of_dxa
(13.) Cosman F, de Beur SJ, LeBoff MS, et al. Clinician's guide to prevention and treatment of osteoporosis. Osteoporos Int. 2014;25: 2359-2381. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4176573/
Table 1. Rates of BMD progression in 391 HIV patients with at least 2 DXAs Progression from normal BMD to osteopenia: 35.7% in median 6.7 y Low risk: Middle risk: High risk: T score > -0.2 T score -0.2 T score -0.6 to -0.6 to -1.0 Median time 8.7 y >7.2 y 1.7 y to progression Progression from osteopenia to osteoporosis: 23.7% in median >8.5 y Low risk: Middle risk: High risk: T score > T score -1.6 T score -2 -1.1 to -1.6 to -2 to -2.4 Median time >8.2 y >8.5 y 3.2 y to progression Source: Negredo et al. (7)
|Printer friendly Cite/link Email Feedback|
|Publication:||Research Initiative/Treatment Action!|
|Date:||Sep 22, 2015|
|Previous Article:||Bone density risk factors in people with HIV: a long, familiar list.|
|Next Article:||Bisphosphonates for bone--and advice on ARVs, calcium, vitamin D, glitazones.|