SAT-515 Racial-Ethnic Differences in Diabetes-Related Fracture Risk

Background: Diabetes mellitus (DM), a novel risk factor for fracture, has not been well studied among African American (AA) and Hispanic (HIS) patients, who have the highest rates of DM in the United States and have higher rates of other major DM complications. Thus, the purpose of this study was to examine DM-related fracture risk among the multi-ethnic population of a large, urban academic center. Methods: We conducted a retrospective study using the electronic medical record. 19,171 subjects with DM (7,622 Caucasians [CA], 7,468 AA and 4,081 HIS) were identified through ICD codes and compared with 26,235 reference subjects with hypertension (HTN) but without DM (15,146 CA, 8,311 AA, and 2,778 HIS). Included subjects were at least 40 years of age, had at least 180 days of follow up, and had at least 1 outpatient visit per year. Subjects with end stage renal disease, bariatric surgery, Paget’s disease, and organ transplantation were excluded. Fractures and co-morbid conditions were identified through ICD codes. Fractures of the humerus, wrist, and hip were analyzed as major osteoporotic fractures (MOF). A Cox proportional hazards model was used to assess the impact of DM on the risk of incident MOF. A random sample of 235 subjects with incident MOF codes was chosen to assess the accuracy of fracture ICD codes and demonstrated an 85% validity rate regardless of race-ethnicity or DM/HTN status. Results: The two cohorts had a total of 717 MOFs over mean 3.2 years of follow-up. Among DM patients, there were significantly more incident MOF in CA and HIS than in AA (unadjusted rates of 1.9% and 1.7% versus 1.1%, respectively, p<0.01 for both). The unadjusted rates of MOF in DM did not differ significantly from rates in the reference cohort with HTN among the same racial-ethnic group (1.8% in CA, 1.3% in HIS, 1.3% in AA). In a comprehensive model that included age, BMI, gender, race, and prior MOF, DM was not significantly associated with the incidence of MOF when the three racial-ethnic groups were analyzed together (HR 1.16, 95% CI 0.99-1.36, p=0.07). However, the effect of DM differed in AA as compared to CA and HIS (AA race by DM interaction p=0.036). DM was significantly associated with the incidence of MOF only in CA and HIS (HR 1.24, 95% CI 1.03-1.49, p=0.02) but not in AA (HR 0.92, 95% CI 0.68-1.23, p=0.56). Interestingly, the association between MOF and prior fracture was far stronger in AA (HR 10.09, 95% CI 5.14-19.81, p<0.001) than in CA and HIS (HR 2.03, 95% CI 1.01-4.09, p=0.048). Conclusion: In a comprehensive model, the presence of diabetes was associated with elevated fracture risk in Hispanics and Caucasians, but not in African Americans. Instead, prior fracture demonstrated a remarkably strong association with incident fracture in African Americans. Our findings may have significant ramifications for case detection and management of diabetes-related fracture risk in multi-ethnic populations in the United States.