Elsevier

The Lancet

Volume 348, Issue 9026, 24 August 1996, Pages 511-514
The Lancet

Articles
Ultrasonographic heel measurements to predict hip fracture in elderly women: the EPIDOS prospective study

https://doi.org/10.1016/S0140-6736(95)11456-4Get rights and content

Summary

Background

The ability of ultrasonographic measurements to discriminate between patients with hip fracture and age-matched controls has until now been tested mainly through cross-sectional studies. We report the results of a prospective study to assess the value of measurements with ultrasound in predicting the risk of hip fracture.

Methods

5662 elderly women (mean age 80 4 years) had both baseline calcaneal ultrasonography measurements and femoral radiography (dual-photon X-ray absorptiometry, DPXA) to assess their bone quality. Follow-up every 4 months enabled us to identify incident fractures. 115 hip fractures were recorded during a mean follow-up duration of 2 years.

Findings

Low calcaneal ultrasonographic variables (obtained from measurements of broadband ultrasound attenuation by, and speed of sound through the bone) were able to predict an increased risk of hip fracture, with similar accuracy to low femoral bone mineral density (BMD) obtained by DPXA. The relative risk of hip fracture for 1 SD reduction was 20 (95% CI 16-24) for ultrasound attenuation and 17 (1 4–2 1) for speed of sound, compared with 19 (16-2 4) for BMD. After control for the femoral neck BMD, ultrasonographic variables remained predictive of hip fracture. The incidence of hip fracture among women with values above the median for both calcaneal ultrasound attenuation and femoral neck BMD was 2 7 per 1000 woman-years, compared with 19 6 per 1000 woman-years for those with values below the median for both measures.

Interpretation

Ultrasonographic measurements of the os calcis predict the risk of hip fracture in elderly women living at home as well as DPXA of the hip does, and the combination of both methods makes possible the identification of women at very high or very low risk of fracture.

Introduction

The lifetime risk of hip fracture for a 50-year-old woman is about 18% and, with increasing life expectancy, a threefold rise in the worldwide incidence of this fracture is expected over the next 60 years.1 The major determinants of hip fracture are intrinsic bone fragility and falls,24 and Melton and colleagues5 have shown that the incidence was inversely related to femoral bone density. Indeed, in one prospective study a decrease of one SD of the hip bone mineral density (BMD) was associated with a doubling of the risk of fractures.6 However, age-related bone demineralisation per se does not completely explain the increasing incidence of hip fracture with age; in people older than 65 the risk of hip fracture, even after adjustment for BMD, approximately doubles each decade.6 Moreover, although patients with hip fracture tend to have a lower BMD than controls, there is a large overlap in BMD between women with fractures and controls matched for age and sex.5, 7 Thus, decreased bone mass may be only one of several pathophysiological factors contributing to bone fragility; an accumulated burden of fatigue damage, and ineffective trabecular architecture, are other possible contributory processes.7, 8

The most commonly used tool for assessing femoral BMD is dual-photon X-ray absorptiometry (DPXA). Although precise, this method provides almost no information about the bone microarchitecture and elasticity. Since microfractures and bone architectural changes may contribute to weakening,7 a combination of information about bone elasticity, structure and density might be a more sensitive predictor of fracture risk than DPXA, which measures only density.

Indirect and in-vitro experience has suggested that ultrasonography may give information not only about the bone density but also about the architecture and elasticity.9, 10, 11 Several cross-sectional studies have shown lower ultrasound measurements for bone quality in osteoporotic (including hip fracture) patients than in controls.12, 13, 14 In 1988, Porter and colleagues15 reported a prospective study of elderly women from institutions, in which broadband ultrasonography of the os calcis showed significantly greater attenuation in hip-fracture patients than in controls, but they did not measure BMD.

The purpose of our study was to assess prospectively, in a large sample of elderly women, whether ultrasonographic measurements of the os calcis were predictive of hip fracture, and whether or not this prediction was independent of femoral-neck BMD.

Section snippets

Patients and methods

7575 women of at least 75 years of age were recruited for the EPIDOS study, a prospective investigation of risk factors for hip fracture conducted in five centres (Amiens, Lyon, Montpellier, Paris, Toulouse) in France. Participants were recruited from registries such as voting or healthcare lists. Women with a history of hip fracture or bilateral hip replacement were excluded; we also excluded women with Paget's disease of bone, malignant bone disease, renal failure, hyperthyroidism, or treated

Results

Because the Achilles device was not available to some centres at the beginning of EPIDOS recruitment, ultrasonographic bone measurement was done for only 5662 of the 7575 women (table 1). During an average of 2 years of follow-up (mean 1·99 [SD 0·57]), 248 women died from unrelated causes and 115 experienced a first hip fracture after minor trauma.

After adjustment for age in elderly women, low values of both calcaneal ultrasonographic measurements and femoral BMD were associated with increased

Discussion

Our results show that ultrasonographic measurements at the calcaneum predict the risk of hip fracture as accurately as do femoral-neck BMD: for each SD decrease in ultrasonographic or densitometric values, the hip-fracture risk is approximately doubled. The magnitude of the relative risk for femoral-neck BMD is consistent with previous reports.4, 5 The only other important prospective study with ultrasonography is that by Porter and colleagues,15 who measured broadband ultrasound attenuation at

References (22)

  • JJ Kaufman et al.

    Ultrasound assessment of bone

    J Bone Miner Res

    (1993)
  • Cited by (0)

    Investigators listed at end of article

    View full text