A comparison of different screening strategies to identify elderly women at high risk of hip fracture: results from the EPIDOS prospective study.

Decision to treat with an anti-osteoporotic drug should be based on individual fracture risk evaluation. We compared the discriminant value of four different screening strategies to identify elderly women with a risk of hip fracture greater than 20 per 1000 woman-years: (1) BMD screening alone, (2) quantitative ultrasound (QUS) screening alone, (3) QUS triage followed by BMD assessment for women with medium-low QUS parameters, and (4) selective BMD screening based on weight followed by clinical evaluation for women with medium-low BMD. The study population included 5910 women aged 75 years or older who participated to the EPIDOS (Epidemiologie de l'Ostéoporose) prospective study. Over an average of 3.7 (+/-0.8) years of follow-up, 231 women suffered a hip fracture, which corresponds to an average risk of 10.6 per 1000 woman-years. All strategies allow us to clearly distinguish a group at high risk of hip fracture (i.e. >20 per 1000 woman-years) from a group at low risk (i.e. below the average risk in the cohort). QUS screening alone has a very low sensitivity (15%). The strategy using QUS as a method of triage and that combining selective BMD and clinical assessment have a sensitivity equivalent to systematic BMD screening (around 35%), with less than 50% BMD examinations. The high-risk women identified by these two strategies are not the same. A simple algorithm combining QUS, BMD, and clinical risk assessment allows an increased number of high-risk women to be identified (21%), and thus improves the sensitivity (53%). With this combined strategy, women in the high-risk group have one chance in ten of having a hip fracture over the next 4 years, whereas women in the low risk group have only one chance in 40.