Background: Densitometric screening for osteoporosis in postmenopausal women has not been demonstrated cost-effective. We have tried to identify clinical factors for screening previous to densitometry avoiding unnecessary explorations.
Setting: outpatient clinics of a menopausal unit in a 450-bed general hospital. Cross-sectional study, in two steps, of two groups of 140 and 284 women attending for physiological menopause. A clinical questionnaire, physical data and lumbar densitometry (Hologic QDR 1000) were obtained classifying the cases as "normal" or "low bone mass" (osteopenia or osteoporosis) according with the WHO criteria. In the first group a logistic regression analysis was done to identify predictive factors for abnormal densitometry, then validated in the second group. Sensitivity, specificity, predictive values (PV) and classification ability of clinical factors were analyzed.
Results: Four factors were independent predictors of abnormal densitometry: age > 51 (odds ratio [OR] = 6.64; 95% CI, 2.36-18.7); body weight < 70 kg (OR = 4.32; 95% CI, 1.71-10.09); years of fertility < 32 (OR = 3.77; 95% CI, 1.36-10.04), and number of live births > 2 (OR = 3.47; 95% CI, 1.27-9.53). Presence of one factor offers: sensitivity 91.9%; specificity 15%; positive PV 66.6%, and negative PV 50%, whereas the presence of two factors offers: sensitivity 62.7%; specificity 70%; positive PV 79.9%, and negative PV 50.3%. Clinical screening allows, when two factors are present, to avoid a 35.5% of densitometries and the false-negative cases represent 18%.
Conclusions: Detection of bone-risk clinical factors (abnormal densitometry) yields a screening, previous to densitometry, that avoids at least one third of explorations in women with physiological menopause, improving the efficiency of the test.