Send to

Choose Destination
See comment in PubMed Commons below
J Clin Densitom. 2007 Oct-Dec;10(4):386-90.

Investigation of the relationship between type 2 diabetes and osteoporosis using Bayesian inference.

Author information

  • 1Department of Epidemiology and Biostatistics, UP College of Public Health, Manila, Philippines.


This study aims to determine the prevalence of Type 2 diabetes in women with osteoporosis and estimate the odds ratio (OR) of osteoporosis in women with Type 2 diabetes using Bayesian inference. This is a case-control study design that looked into prevalence of diabetes among 582 female patients who had normal bone mineral density (BMD) and 598 female patients with osteoporosis. The subjects included women at least 30 yr of age who had their BMD measured in the lumbar spine and femoral neck using dual-energy X-ray absorptiometry at a tertiary referral center in Manila, Philippines. Prevalence of Type 2 diabetes in subjects with osteoporosis is 22.41%, whereas 19.07% of the subjects with normal BMD had diabetes. The odds of developing osteoporosis is 22.54% higher for Type 2 diabetic subjects. Patients with osteoporosis were older than subjects with normal BMD by almost 10 yr. Of the diabetic osteoporotic patients, 44.78% were physically active compared with 20.72% diabetics with normal BMD. Most of the diabetics (60.36%) with normal BMD were obese, whereas majority of diabetic osteoporotics (64.93%) have normal body mass index (BMI). Less than 10% of both diabetic osteoporotics and diabetics with normal BMD have ever undergone hormone replacement therapy. Of the 598 subjects with osteoporosis, 124 (20.74%) had suffered from fragility fractures. When controlling for physical activity and BMI, the odds of developing osteoporosis was 21.73% and 53.89% higher for Type 2 diabetics, respectively. In considering all possible confounders and effect modifiers (age, physical activity, BMI, and hormone replacement therapy) in the model which made use of a diffuse normal prior distribution, the estimate for OR (Model 1) is 0.67. A separate analysis excluding modifiable confounders (Model 2) gave the measure of association an equal likelihood of diabetes being a protective factor or a risk factor. The crude OR indicated that Type 2 diabetes is a risk factor for osteoporosis. However, when identified confounders were included in the model, the direction of the relationship changed. Considering the credible intervals (95% credible interval in both models), the study concluded that diabetes is indeed a protective factor for osteoporosis. Results of the study may have potential limitations. There are sources of bias that have been identified--selection bias where patients included in the study were referred by primary care givers for a specified reason as well as misclassification and recall biases on certain information such as type and duration of physical activity. Diabetes is a protective factor for osteoporosis in this referred population of women. However, with the well-known diabetes-related factors, that is, microvascular complications, visual acuity, and risk for fall, one should still strongly consider assessing and screening for osteoporosis and fracture risk reduction in diabetic patients.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Elsevier Science
    Loading ...
    Write to the Help Desk