Bone Mineral Density, Pioglitazone Use, Vitamin D, Androgen, and Estrogen Levels in an Inner City Type 2 Diabetes Mellitus Population; Part of a Cross-Sectional Study

Presentation Number: SAT-259
Date of Presentation: June 15th, 2013

Marianna Antonopoulou*1, Thomas Nithin2, Gul Bahtiyar2, Samy I McFarlane3 and Alan Scott Sacerdote4
1SUNY Downstate Medical Center, Brooklyn, NY, 2Woodhull Medical Center, Brooklyn, NY, 3SUNY Health Sci Ctr-Brooklyn, Brooklyn, NY, 4Woodhull Med & Mental Hlth Ctr, Brooklyn, NY

Abstract

Pioglitazone is useful in improving glycemic control and improving several aspects of the metabolic syndrome, but its use is concerning regarding bone density (BMD), fracture risk, weight gain, edema, and heart failure. Most published data on BMD and fracture risk is on white patients.

From 2007-2012 we studied 218 patients with Type 2 Diabetes (Type 2 DM), 166 women, 52 men in a cross-sectional design via chart review. Patients were 65.6% Hispanic, 28% Black, 4.6% White, and 1.8% Asian. 49.3% were taking pioglitazone (P). Mean duration of diabetes was 7.76+1.96 years. The two groups were quite comparable in terms of age, BMI, A1c, Vitamin D, and PTH levels. We compared absolute mean BMD and t-score at the lumbar spine, total hip and femoral neck in the 2 groups by unpaired t-test and then corrected for possibly confounding variables by multiple regression analysis using SPSS.

After correction, absolute mean BMD was significantly lower in P users only at the femoral neck p=0.16. Black patients had significantly higher mean BMD for the total hip than patients from other backgrounds p=0.02. Serum estrone was significantly, negatively correlated with patient age p=0.006. Serum estradiol was significantly correlated with patient BMI p=0.01. Serum androstenedione was significantly, negatively correlated with patient age p=0.038. Serum total and free testosterone were significantly, negatively correlated with patient BMI, p=0.12, p=0.004 respectively. Sex hormone binding globulin was significantly, negatively correlated with HbA1c p=0.035. T-score at total hip was significantly lower in P users p=0.015, was negatively correlated with patient age p=0.009, positively correlated with BMI p<0.001, and was higher in Black patients than in other groups p=0.001. T-score in the lumbar spine was significantly lower in P users p=0.026 and was higher in Black patients p=0.01.

 

  • In this largely Hispanic and Black cohort P use was associated with modestly lower BMD that may be mitigated by concurrent use of metformin, insulin, hydrochlorothiazide, and/or sulfonylureas, and offset by increasing BMI.
  • The observed mitigating effect of increasing BMI may be partially explained by the association of BMI with estradiol.
  • Black patients are relatively protected in terms of absolute BMD at total hip and t-score at the lumbar spine.

 

Nothing to Disclose: MA, TN, GB, SIM, ASS