Skip navigation

Matters postmenopausal

Bandolier has written before ( 25 , 37 ) about hip fractures, and how devastating these are for people who suffer them. Elderly persons with a hip fracture are unlikely to regain their independence, and some degree of permanent disability is probable. Hips and other bones will fracture for different reasons, but osteoporosis will be a major factor in the elderly, especially women.

Endogenous hormones

Oestrogen has important effects on bone metabolism. Lack of it increases the risk of bone resorption together with loss of mineral results in bones getting progressively weaker. The importance of oestrogen is shown by a study on how endogenous oestrogen affects the risk of hip and vertebral fractures in older women [1].


Between 1986 and 1988 just under 10,000 women aged 65 years or more were recruited in four US cities. They were asked about lifestyles, oestrogen replacement, and multivitamin and calcium use. X-rays of the spine and bone mineral density measurements were taken, together with blood samples for measurement of a variety of hormones. The women were then contacted by mail every four months to identify fractures. Follow up was more than 99% complete. Follow up spine X-rays were obtained in the 79% of the women still alive at 3.7 years.

A random selection was made of 133 of the 332 women who had a hip fracture and 138 of the 359 women with a new vertebral fracture, together with control groups randomly selected and twice to three times the number.


Women were well matched at baseline for calcium and vitamin D supplementation and intake, though women with fractures were older, lighter and had lower bone density. After adjustment for weight and age two hormones were particularly associated with increased risk of fracture of the hip or vertebrae.

A serum oestradiol that was undetectable (<18 pmol/L) was associated with a relative risk of about 2.5 (1.4 to 4.4). Any increase in serum oestradiol above the limits of detection slightly more than halved the risk of a fracture (Figure).

A serum sex hormone binding globulin of ≥10 μg/L was associated with a relative risk of 2.1 (1.1 to 4.2). Any increase in sex hormone binding globulin above 10 μg/L resulted in progressive increases in the risk of fracture, up to three-fold at the highest concentrations found (Figure).

Women with undetectable levels of oestradiol and measurable sex hormone binding globulin had a much increased risk of fracture, even when adjustments were made for age and weight (Table).

Relative risk (95% CI) for fracture in women with undetectable levels of oestradiol and measurable sex hormone binding globulin
  Hip fracture Vertebral fracture
Age adjusted 14 (3.0 to 62) 12 (3.3 to 41)
Weight adjusted 6.9 (1.5 to 32) 7.9 (2.2 to 28)


Sex hormone binding globulin binds oestradiol into a form that is not immediately available to tissues. Both increases in sex hormone binding globulin and low levels of oestradiol work in the same way - to starve tissues, including bone, of available oestrogen. This important study dramatically highlights the importance of oestrogen as a risk factor for hip and vertebral fractures. If the associations were causal, they would account for a substantial proportion of fractures in elderly women. There may even be a suggestion that women particularly at risk could be identified by blood tests, and remedial therapy instituted. Bandolier does not leap to the conclusion that screening is the answer, or an answer, but there is an opportunity here to further explore this interesting area. It could be one of those topics where diagnostic tests and treatment together make good sense.

These results also give some biological plausibility for the effects of Soya ( Bandolier 56 ). Soya isoflavenoids are weakly oestrogenic, and probably provide a low level of oestrogenic 'cover'. Whether they have effects on hip and vertebral fractures still has to be clarified.


  1. SR Cummings et al. Endogenous hormones and the risk of hip and vertebral fractures among older women. New England Journal of Medicine 1997 339: 733-738.

previous or next story in this issue