OBJECTIVES: to assess the main risk factors associated to low bone mass and to propose algorithms for the indication of the evaluation of bone mineral density in a women population from São José do Rio Pardo, state of São Paulo. METHODS: 324 women (mean age 60 years old) were studied. A questionnaire was sent by regular mail to collect data for several osteoporosis risk factors, as well as both personal and mother’s history of fractures. Bone mineral density was measured using DXA, measuring the lumbar spine segment (L2-L4), and/or the proximal femur (femoral neck). Data were classified according to the WHO criteria. Data were analyzed using one-way Anova, and the "post-hoc" Student-Newman-Keuls test for paired multiple comparisons; 2 x2 and k x r contingency tables and chi-square test were also used. The most efficient variables for the identification of women at risk for osteoporosis were determined through multiple regression analysis. The previously described algorithms Osteorisk and Osiris were also applied. RESULTS: there were 88 (27.2%), 160 (49.4%), and 76 (49.4%) women among the normal, osteopenia, and osteoporosis groups, respectively. Mean age was 56, 60, and 65 years-old for the normal, osteopenia, and osteoporosis groups, respectively. Mean weight and height was 69.5 kg/1.60 m, 64.9 kg/1.59 m and 62.1 kg/1.57 m for the normal, osteopenia, and osteoporosis groups, respectively. Differences between these mean values were statistically significant. The number of years of menopause was higher in the osteoporosis group; contraceptive and hormone replacement therapy was less frequent in the osteoporosis group, as well as menopause. All women in the osteoporosis group were post-menopause, compared to 86% in the other groups. Women in the osteoporosis group had more previous fractures, reaching statistical significance. Using multiple regression, based on the main variables identified, we reached a formula considering age, weight, and height as the osteoporosis predicting model that was compared to the two algorithms Osteorisk and Osiris. CONCLUSION: among the variables studied, age and weight showed higher significance. Their combination in an algorithm showed better performance than either of the variables isolated. The inclusion of new variables did not increase the performance of the identification of osteoporotic women. The use of an algorithm with objective criteria makes it possible to better select candidates for bone densitometry analysis.
osteoporosis; risk factors; bone mineral density; female-gender
