Background/Purpose: Fractures are potentially avoidable and highly morbid. We studied non-vertebral fracture rates among SLE patients enrolled in Medicaid, the U.S. health insurance program for low-income people, and compared them to age- and sex-matched Medicaid patients without SLE.

Methods: Among patients aged 5-65 years enrolled >180 days during 2007-2010, we identified an SLE and a lupus nephritis (LN) cohort using previously described algorithms (SLE: ≥3 ICD-9 codes for SLE ≥30 days apart; LN: SLE + ≥2 ICD-9 codes for kidney disease ≥30 days apart). Four age- and sex-matched non-SLE patients were included for each SLE patient. Subjects were followed from index date (SLE definition met, or a matched date for non-SLE patients) through end of follow-up (12/31/10) or censoring (death or disenrollment). During the baseline period (180 days before index date), we assessed sociodemographics, SLE- and prescriptions, and comorbidities. Incident fractures of the pelvis, hip, wrist, and humerus were identified via validated algorithms using ICD-9 + current procedural terminology codes. Within each cohort, we calculated cumulative incidence rates (IRs) and 95% confidence intervals (CI) for any fracture and for each fracture type. Poisson regression models, adjusted for all covariates, estimated cumulative IR ratios (IRRs), comparing SLE and LN to non-SLE patients. Age-stratified analyses investigated differences in fracture rates, using age 45 years as a cutoff based on mean age at menopause in SLE.

Results: 39,918 SLE patients were matched to 159,672 non-SLE patients. Mean age was 40.1 (±13.4) years and 92.2% were female. 9,096 SLE patients (22.8%) had LN. Black race comprised 43.3% of SLE, 53.1% of LN, and 23.1% of the non-SLE cohort. Average prednisone equivalent during baseline was ≥7.5mg/day in 11.5% of SLE, 20.1% of LN, and 0.3% of non-SLE patients. Fracture IRs were 1.48/100,000 person-years in SLE (Table) and 2.06/100,000 person-years in LN. Compared to matched non-SLE patients, fracture rates were 39% higher in SLE patients and 69% higher in LN patients, even after adjusting for medications associated with bone density loss. Hip fracture rates were almost 3.5 times higher and pelvic fracture rates were more than doubled among SLE patients ≤age 45 compared to matched non-SLE patients after adjustment. Hip fracture rates were 5 times higher among LN patients ≤ age 45 compared to matched non-SLE patients (adjusted IRR 5.02 [1.46-17.19]).

Conclusion: Compared to age- and sex-matched Medicaid enrollees without SLE, SLE and LN patients had much higher non-vertebral fracture rates, particularly hip and pelvic fractures in patients ≤ 45 years of age. These results highlight the need for fracture prevention measures among young SLE patients.