Date of release: 24 July, 2017

FRAX^®

Introduction

Since its launch in 2008, the web-based fracture risk assessment tool FRAX^® [1] has been evaluated thoroughly in additional validation studies and widely published in meta-analyses and clinical review papers. Furthermore, its predictive and discriminative powers have been compared with other osteoporotic fracture risk prediction tools [2]. Two recent papers elucidate different aspects of the tool, one assessing its diagnostic accuracy in women and men from five different non-US populations [3], the other one calculating the time to 'clinically relevant' risk scores in US postmenopausal women [4], and both using the 10-year intervention thresholds of 20% for major osteoporotic fractures (MOF) and 3% for hip fractures (HF), as suggested by the National Osteoporosis Foundation (NOF) [5].

In the first paper, a systematic review and meta-analysis of seven studies from New Zealand, Canada, the USA, France and Poland, which tested FRAX^® in populations other than the derivation cohorts, the tool 'performed better in identifying patients who will not have a MOF or HF within 10 years, than those who will. A substantial number of patients who developed fractures, especially MOF within 10 years of follow-up, were missed by the baseline FRAX^® assessment', as stated by the authors in their conclusion.

For MOF prediction, the mean sensitivity, specificity, and diagnostic odds ratio (DOR) along with their 95% confidence intervals (CI) were 10.25% (3.76–25.06%), 97.02% (91.17–99.03%) and 3.71 (2.73–5.05); for HF prediction 45.70% (24.88–68.13%), 84.70% (76.41–90.44%) and 4.66 (2.39–9.08), respectively, the latter one being less precise because of its larger confidence region.

FRAX^® is freely available and easy to use, not least because of its condensed and time-saving features. But this is at the expense of its sensitivity. Tools with a larger number of clinical risk factors, e.g. QFracture^®, may be more sensitive but less feasible [2]. Lowering the intervention threshold may also improve sensitivity but increase over-treatment [6].

The second paper estimates the timing of occurring ‘clinically relevant’ scores, i.e. treatment-level FRAX^® scores, according to 2014 National Osteoporosis Foundation guidelines [5], and screening-level FRAX^® scores, according to 2011 US Preventive Services Task Force (USPSTF) guidelines [7], in postmenopausal women of the Women’s Health Initiative (WHI) cohort.

The screening level was proposed by the USPSTF to identify postmenopausal women < 65 years of age as candidates for bone mineral density (BMD) testing by dual-energy X-ray absorptiometry (DXA) if their FRAX^® score was equal to or greater than that of a 65-year-old white woman who has no additional risk factors, i.e. a 10-year risk for MOF of ≥ 9.3%. The screening-level score had low predictive ability in postmenopausal women aged 50–64, as shown before with data of the same cohort and in comparison to other non-satisfying fracture risk tools [8], leading to the conclusion '… that fracture prediction in younger postmenopausal women requires assessment of risk factors not included in currently available strategies.' The treatment-level score without BMD measurements occurred in 10% of women aged ≥ 65 after 3–5 years of baseline score (effective maximum length of follow-up 18.6 years), and the treatment-level score with BMD measurements after 5–7 years (maximum follow-up 11.2 years), respectively. Thus, scores calculated without BMD indicated higher risk than scores with BMD.

The authors conclude by confirming a familiar sounding message: 'Postmenopausal women with sub-threshold fracture risk scores at baseline were unlikely to develop a treatment-level FRAX^® score between ages 50 and 64. After age 65, the increased incidence of treatment-level fracture risk scores …. supports more frequent consideration of FRAX^® and bone mineral density (BMD) testing.'

Nevertheless, we are confronted with a high prevalence of osteoporosis and fragility fractures in middle-aged women (< 65 years) [9,10] and badly need to improve our strategies of identifying women at risk as early as age 40 by including more clinical risk factors, e.g. detrimental gyneco-endocrinological conditions such as premature ovarian insufficiency, laboratory tests as bone turnover markers, and, last but not least, intensified interdisciplinary cooperation in preventive measures.