OBJECTIVE: To determine risk factors for osteoporosis in postmenopausal women from Mexico City. METHOD: The following prognostic variables for low bone density were analyzed in 400 postmenopausal women: age, body mass index (BMI), time since menopause (TM), previous and current hormone replacement therapy (HRT), previous use of corticosteroids, thyroid disease, and previous fractures. All patients were submitted to bone absorptiometry (DEXA) considering the diagnostic criteria for osteoporosis established by the WHO. All variables were analyzed in a bivariate mode, and logistic regression analysis was performed for variables that proved significant (age, TM, and BMI), estimating odds ratios and 95% confidence intervals. RESULTS: Age, TM and BMI resulted as significant risk factors for osteoporosis (P < .01), with the following cut-off points: age < or = 48 and >48 years; TM < or = 5 and >5 years; and BMI < or = 32 and >32 for L1-L4 average, BMI < or = 30 and >30 for femoral neck (FN), < or = 27 and >27 for trochanter, and < or = 32 and >32 for Ward's triangle (WT). Odds ratio (OR) for age was 1.57 at L1-L4 average (95% CI 1.01-2.47); OR for TM was 2.09 at L1-L4 average (95% CI 1.24-3.55), 2.05 at FN (95% CI 1.23-3.47) and 1.78 in WT (95% CI 1.05-2.99). OR for BMI was 0.36 at L1-L4 average (95% CI 0.17-0.80), 0.25 at FN (95% CI 0.12-0.5), 0.18 at trochanter (95% CI 0.10-0.32) and 0.29 at WT (95% CI 0.13-0.63). CONCLUSIONS: The various risk factors exerted different influences on the bone areas analyzed. Age increased the risk for lumbar spine osteoporosis, TM had a deleterious effect, while BMI showed a beneficial effect on bone density. These factors should be considered to establish the risk of osteoporosis in postmenopausal women, and to decide which patients should be submitted to absorptiometry.