Factors Associated With Low Bone Mineral Density in Adults With Type 1 Diabetes: A Cross-Sectional Study

Abstract

Context: Individuals with type 1 diabetes (T1D) have a two- to threefold increase in fracture risk at any site, and up to a sevenfold increase in hip fracture risk compared to those without diabetes. The mechanisms accounting for this bone fragility are not yet fully understood. Objectives: 1) To determine factors associated with low bone mineral density (BMD) in patients with T1D; 2) To assess the association between skin advanced glycation end products (AGEs) and low BMD in patients with T1D. Methods: These are preliminary data from patients with T1D included in a cross-sectional study aiming at comparing the prevalence of vertebral fractures between adult patients with T1D from two tertiary care centers and age- and sex-matched controls without diabetes. Patients were eligible if they were aged ≥20 years and had a diagnosis of T1D for at least 5 years. Patients were classified as having a low BMD if Z-score was ≤-2.0 at any site (lumbar spine, femoral neck, total hip, radius) in patients aged <50 years or if T-score was ≤-1.0 at any site in patients aged ≥50 years or in postmenopausal women. Skin AGEs (surrogate marker of overall including bone AGEs) were measured by skin autofluorescence (AGE Reader ®). Unpaired t-tests or Chi-squared tests were used to compare characteristics between patients with or without a low BMD. Variables associated with a low BMD were determined by univariate analysis and were subsequently included in a multivariate logistic regression analysis if p<0.1 in the univariate analysis. All variables were tested for multicollinearity. Results: 106 patients with T1D were included (mean age 45.2±15.0 years; mean BMI 26.3±5.1 kg/m2; 54.7% women; mean duration of diabetes 28.2±13.6 years; 44.3% with a microvascular complication). Mean HbA1C over the past 3 years was 7.5±0.8%. A third of the patients (31.1%) had a low BMD (3 patients using Z-score; 30 patients using T-score). Patients with a low BMD were older (58.3 vs 39.3 years, p<0.001), had a lower mean HbA1C over the past 3 years (7.3% vs 7.6%, p=0.047), a longer diabetes duration (36.1 vs 24.6 years, p<0.001), higher skin AGEs (2.50 vs 2.03, p<0.001), a higher prevalence of microvascular complications (63.6% vs 37.7%, p=0.02) and a higher prevalence of abnormal albumin to creatinine ratio (ACR ≥2.0) on the day of assessment (38.7% vs 11.8%, p=0.003). In multivariate regression analysis, age (p<0.001), abnormal ACR (p=0.003) and lower mean HbA1C over the past 3 years (p=0.02) remained significantly associated with a low BMD. Skin AGEs were correlated with age (r=0.56) and diabetes duration (r=0.47). Conclusion: In this population with T1D, a low BMD was independently associated with older age, abnormal ACR and, unexpectedly, with a lower mean HbA1C over the past 3 years, but not with skin AGEs.