Research Article

Prediction of hip fracture in elderly women: a prospective study.

BMJ 1990; 301 doi: (Published 29 September 1990) Cite this as: BMJ 1990;301:638
  1. R W Porter,
  2. C G Miller,
  3. D Grainger,
  4. S B Palmer
  1. Doncaster Royal Infirmary.


    OBJECTIVE--To assess the relative importance of osteoporosis of the os calcis, cognisance, and mobility in the risk of subsequent fracture of the hip in elderly women. DESIGN--Prospective study of elderly women in residential care over two years. SETTING--21 Private or 38 local authority residential homes for the elderly and 4 geriatric hospitals in Doncaster and Hull. SUBJECTS--1414 Ambulant women aged over 69, in private or local authority residential care or geriatric care. Those who had had bilateral hip surgery were excluded. MAIN OUTCOME MEASURES--Broad band ultrasonic attenuation (BUA) index, Clifton assessment procedures for the elderly test (for cognisance), and mobility on a six point scale, and fracture of the hip in the subsequent two year period. RESULTS--73 Women fractured their hip during the two years. Their mean age was not significantly different from that of the women who did not have a fracture (85.3 (SD 5.6) v 83.9 (6.3); p = 0.07), but their mean BUA index (40.3 (19.3) v 50.9 (22.2) db/MH2), and score for cognisance (median 19 (interquartile range 10.5-27.0) v 24 (17-30)) were significantly lower (both p less than 0.001). These variables had independent associations with fracture of the hip. Women with fractures had a significantly lower score for the psychomotor component of the cognisance test (4.5 (1-8) v 7 (2-10); p less than 0.0025 and were significantly more mobile (1(1-3) v 3 (1-6); p less than 0.02). Subdividing women according to high, medium, and low scores for BUA index and cognisance testing disclosed a high risk group (118 women) with low BUA index and cognisance score, whose incidence of fracture was 12.8%; in the group at lowest risk (136 women) with high BUA index and cognisance score, the incidence of fracture was only 1.5% (relative risk 8.4 (95% confidence interval -2.0 to 35.5]. Further analysis showed that those most at risk were, additionally, most mobile but that less mobile women with good cognisance had a low incidence of fractures, regardless of the BUA index, (1.2%, high index, v 0.9%, low index). CONCLUSIONS--Elderly women most at risk of sustaining hip fractures were those with low BUA index, low cognisance test score, and high mobility. Improving bone strength and cognisance in elderly women may reduce their incidence of hip fracture.