Predictors of height loss in postmenopausal women
Age, weight, oral corticosteroid use and strenuous exercise at age 18 years appear to be useful in predicting the 5-year risk of height loss of at least 1 inch in postmenopausal women, according to a study.
Researchers measured height in 1,024 postmenopausal women (mean age 65.9 years) at baseline and at 5-year follow-up using a fixed stadiometer. Demographics, lifestyle, medical history and medication use were also assessed at baseline.
A mean loss of 0.4 inches in height occurred during an average follow-up of 5.1 years. Height loss was at least 1 inch in 70 women (6.8 percent). Between women who lost ≥1 inch in height and those whose height loss was <1 inch, there were significant differences in age, weight at baseline, waist/hip ratio, body mass index, alcohol intake, years since menopause, diagnosed/treated for osteoarthritis, baseline use of oral corticosteroids, oestrogen therapy, energy expenditure (MET-h/wk) from total exercise and forearm bone mineral density (BMD; p<0.05 for all).
On multivariable stepwise logistic regression analysis, height loss ≥1 inch was significantly associated with age (odds ratio [OR], 1.11; 95 percent CI, 1.06–1.16), weight (OR, 1.05; 1.03–1.07), use of oral corticosteroids (OR, 4.96; 1.25–19.72) and strenuous exercise ≥3 times per week at age 18 years (OR, 0.55; 0.31–0.98).
A prediction model consisting of the factors stated above was able to differentiate women who had marked height loss from those who did not (area under the receiver operating characteristic curve, 72.1 percent). The addition of BMD measures did not improve the discriminatory ability of the prediction model.
Researchers pointed out that the current prediction model has potential utility in identifying women at risk of height loss where interventions may be beneficial. More important, it is likely that earlier identification and treatment of these women may attenuate the effect of height loss on outcomes such as fracture and mortality in elderly women.
However, further studies are needed to validate the current model in other independent populations before it can be used in primary care settings, researchers added.