Body mass index and bone loss among postmenopausal women: the 10-year follow-up of the OSTPRE cohort

Abstract

Obesity protects against osteoporosis, but the magnitude of this association has been difficult to assess from cross-sectional or short term studies. We examined the time course of bone loss as a function of body mass index (BMI) in early and late postmenopausal women. Our study population (n = 300) was a random sample of the population-based Kuopio Osteoporosis Risk Factor and Prevention (OSTPRE) Study, Finland. We excluded women without complete BMD results, premenopausal women during the second bone densitometry and women who had used hormone replacement therapy, bisphosphonates or calcitonin. BMI along with femoral neck and spinal bone mineral density (BMD) were assessed three times by dual-energy X-ray absorptiometry during a mean follow-up of 10.5 years (SD 0.5). The mean baseline age was 53.6 years (SD 2.8), time since menopause 2.9 years (SD 4.3) and BMI 27.3 kg/m² (SD 4.4). The data was analyzed by linear mixed models. Thus, we were able to approximate the bone loss up to 20 postmenopausal years. To illustrate, a woman with a baseline BMI of 20 kg/m² became osteopenic 2 (spine) and 4 (femoral neck) years after menopause, while obesity (BMI of 30 kg/m²) delayed the incidence of osteopenia by 5 (spine) and 9 (femoral neck) years, respectively. The delay was due to high baseline BMD of the obese, while bone loss rate was similar for both lean and obese subjects. This lean versus obese difference may also be partly due to altered X-ray attenuation due to fat mass.

Appendix 1

Equations for estimating postmenopausal (0–20 years) and age-related (50–70 years) change in bone mineral measurement values (n = 300) as a function of body mass index (BMI). Bone mineral measurements and all anthropometrical parameters were measured three times during a mean follow-up of 10.5 years. Statistical method: mixed models framework in the form of random coefficients regression (see Materials and methods for details). Equations for estimating age-related and postmenopausal changes in weight and height are also shown. All factors in the models were statistically significant (p < 0.05). BMD (g/cm²) i.e., bone mineral density.

Equations according to time since menopause (TSM, years):

$$ \begin{gathered} {\text{Lumbar spine BMD}} = - 0.0131 \times {\text{TSM}} + 0.0005 \, \times {\text{TSM}}^{2} + 0.0037 \times {\text{BMI}} + 1.0421 \hfill \\ {\text{Femoral neck BMD}} = - 0.0079 \times {\text{TSM}} + 0.0069 \times {\text{BMI}} + 0.7616 \hfill \\ \end{gathered} $$

$$ \begin{gathered} {\text{Lumbar spine T-score}} = - 0.1061 \times {\text{TSM}} + 0.0040 \times {\text{TSM}}^{2} + 0.0298 \times {\text{BMI}} - 1.3566 \hfill \\ {\text{Femoral neck T-score}} = - 0.0559 \times {\text{TSM}} + 0.0492 \times {\text{BMI}} - 1.7547 \hfill \\ \end{gathered} $$

$$ \begin{gathered} {\text{Weight}} = 0.4624 \times {\text{TSM}} - 0.0078 \times {\text{TSM}}^{2} + 69.8924 \hfill \\ {\text{Height}} = - 0.1241 \times {\text{TSM}} + 161.3485 \hfill \\ \end{gathered} $$

Equations according to calendar age (years):

$$ \begin{gathered} {\text{Lumbar spine BMD}} = - 0.0812 \times {\text{age}} + 0.0006 \times {\text{age}}^{2} + 0.0047 \times {\text{BMI}} + 3.4827 \hfill \\ {\text{Femoral neck BMD}} = - 0.0269 \times {\text{age}} + 0.0002 \times {\text{age}}^{2} + 0.0072 \times {\text{BMI}} + 1.7157 \hfill \\ \end{gathered} $$

$$ \begin{gathered} {\text{Lumbar spine T-score}} = - 0.6604 \times {\text{age}} + 0.0052 \times {\text{age}}^{2} + 0.0383 \times {\text{BMI}} + 18.4853 \hfill \\ {\text{Femoral neck T-score}} = - 0.1893 \times {\text{age}} + 0.0011 \times {\text{age}}^{2} + 0.0506 \times {\text{BMI}} + 4.9768 \hfill \\ \end{gathered} .$$

$$ \begin{gathered} {\text{Weight}} = 2.0041 \times {\text{age}} - 0.0142 \times {\text{age}}^{2} + 4.3970 \hfill \\ {\text{Height}} = - 0.1246 \times {\text{age}} + 167.6488. \hfill \\ \end{gathered} $$