NPC Archive Item: Simple models based on age and BMD predict 10-year risk of fracture as well as more complex models

NOTE – This is an archive post from the NPC and has not been updated since first publication. Therefore, some hyperlinks may no longer be working.
MeReC Rapid Review NPC Logo

3 February 2010

A study has shown that, in older community-dwelling women, a simple model based on age and bone mineral density alone (BMD), such as that recommended by NICE (TA160 and TA161), predicted 10-year risk of hip, major osteoporotic, and any clinical fracture as well as more complex FRAX® models with BMD.

Level of evidence:
Level 2 (limited quality patient-oriented evidence) according to the SORT criteria.

Action
Healthcare professionals should continue to follow recommendations in the two NICE technology appraisals on drug treatments for primary and secondary prevention of osteoporotic fragility fractures in postmenopausal women. NICE used detailed modelling of the cost-effectiveness of each therapy at various stages before making treatment recommendations for a number of groups. They concluded that it is appropriate to recommend treatment on the basis of a combination of T-score, age and the number of independent clinical risk factors for fracture. T-score relates to the measurement of BMD using hip and/or spine dual-energy X-ray absorptiometry (DXA) scanning and is calculated by comparing the observed BMD with that of an average healthy young premenopausal woman.

What is the background to this?
It is important to distinguish between risk of osteoporosis and risk of osteoporotic fracture. As well as increasing age and low BMD many other factors can determine the risk of fracture. NICE divided them into two groups: firstly, those that independently raise fracture risk (parental history of hip fracture, alcohol intake of four or more units a day, prior fracture, rheumatoid arthritis [RA]); and secondly, indicators of low BMD (low body mass index [BMI], certain medical conditions, prolonged immobility, untreated premature menopause). Other issues may impact on the risk of falls or fractures, for example, calcium and vitamin D deficiency, physical activity level, muscle strength, maintenance of balance while upright, visual impairment and use of sedating medications.

Estimating the risk of fracture using an individual’s risk characteristics is inaccurate. Therefore, as with cardiovascular disease, a risk calculator (FRAX®) was developed by the World Health Organization including a wide range of risk factors (age, sex, fracture history, use of oral glucocorticoids, presence of RA and other conditions, parental history of hip fracture, smoking status, alcohol intake, and BMI). NICE has not assessed the use of FRAX® and their recommendations are not based on this risk calculator for a number of reasons. Firstly, not all the risk factors it includes may be appropriate, e.g. smoking. Secondly, absolute risk of fracture is not an accurate predictor of the cost effectiveness of fracture prevention drugs because different fracture sites have different impacts on quality of life, cost and mortality. Thirdly, the evidence base does not currently show that modifying all of the risk factors included in the calculator results in clinical benefit.

This study used data collected in the Study of Osteoporotic Fractures to compare the value of FRAX® models with femoral neck BMD with that of simple models based on age and BMD alone for prediction of fractures in over 6,000 women aged 65 years or older. In addition, because BMD is not always available, a secondary analysis was performed comparing FRAX® models without BMD with simple models based on age and fracture history alone. Women were followed up for 10 years to ascertain the risk of hip, major osteoporotic (hip, clinical vertebral, wrist or humerus) and any clinical fracture.

What does this study claim?
The study concluded that a simple model based on age and BMD alone predicted 10-year risk of hip, major osteoporotic, and any clinical fracture as well as more complex FRAX® models with BMD. Similarly, a parsimonious model based on age and fracture history alone predicted 10-year risk of these three fracture outcomes as well as more complex FRAX® models without BMD.

These findings suggest that use of the FRAX® risk assessment tool does not enhance fracture prediction in older women beyond that provided by simple models based on age and BMD or age and fracture history alone. An accompanying commentary points out only 21.8% of older women with a prior fracture are tested or treated for osteoporosis. The wider adoption of evidence-based practice in this group may be a more useful activity pending definitive research data being available, rather than debating the pros and cons of various ways of identifying risk,

How does this relate to other studies?
Another study has described the development and validation of two new fracture clinical risk scores (QFractureScores™) for men and women representative of the primary care population in England and Wales.

Routinely collected data from over one million men and over one million women aged 30 to 85 years from 357 general practices in England and Wales were examined to develop the scores. The two primary outcome measures were first diagnosis of osteoporotic fracture (vertebral, distal radius, or hip) and incident hip fracture recorded in general practice computer records. Data collected from a further 178 practices was used to validate the scores. Performance of the QFractureScore™ in predicting the risk of hip fracture was compared with FRAX®.

The authors concluded that the validation statistics for the hip fracture algorithm suggest that the models are likely to be at least as effective at identifying patients at high risk of hip fracture within primary care as the FRAX® algorithm. However, further validation studies are needed to test the performance of these algorithms in independent populations that are representative of the setting where the algorithms are likely to be used. In addition, as we saw in the first study, simple models based on age and BMD alone, or age and fracture history alone, predicted 10-year risk of hip, major osteoporotic, and clinical fracture as well as more the complex FRAX® models.

Healthcare professionals should continue to recommend treatment to prevent osteoporotic fragility fracture in postmenopausal women according to NICE guidance i.e. on the basis of a simple combination of T-score, age and number of independent clinical risk factors for fracture.

Further information on the management of osteoporosis is available in a suite of materials on the osteoporosis section of NPC. In addition a recent MeReC Bulletin illustrates use of the two NICE technology appraisals on drug treatments for primary and secondary prevention of osteoporotic fragility fractures in postmenopausal women, and outlines some issues for healthcare professionals to consider.

Study details
Ensrud KE, Lui L-Y, Taylor BC, et al. A comparison of prediction models for fractures in older women. Is more better? Arch Int Med 2009;169:2087–94

Design: Prospective cohort study.

Patients: 6,252 women aged 65 years or older.

Intervention and comparison: The value of FRAX® models that include BMD was compared with that of parsimonious models based on age and BMD alone for prediction of fractures. In addition FRAX models without BMD were compared with simple models based on age and fracture history alone.

Outcomes: Hip, major osteoporotic (hip, clinical vertebral, wrist, or humerus), and any clinical fractures were ascertained during 10 years of follow-up. Area under the curve (AUC) statistics from receiver operating characteristic curve analysis were compared between FRAX® models and simple models.

Results: The AUC comparisons showed no differences between FRAX® models with BMD and simple models with age and BMD alone in discriminating hip (AUC 0.75 for the FRAX® model and 0.76 for the simple model; P=0.26), major osteoporotic (AUC 0.68 for the FRAX® model and 0.69 for the simple model; P=0.51), and clinical fracture (AUC 0.64 for the FRAX® model and 0.63 for the simple model; P=0.16). Similarly, performance of parsimonious models containing age and fracture history alone was nearly identical to that of FRAX® models without BMD. The proportion of women in each quartile of predicted risk who actually experienced a fracture outcome did not differ between FRAX® and simple models (P 0.16).

Sponsorship: The Study of Osteoporotic Fractures is supported by National Institutes of Health funding.

Please comment on this blog in the NPC discussion rooms, or using our feedback form.

Make sure you are signed up to NPC EMail updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you