Abstract
OBJECTIVE: We aimed at determining the predictors of osteonecrosis (ON) in a longitudinal lupus cohort. METHODS: Data were reviewed from the initiation of the cohort in 1987 until October 2019. In total, 2,428 patients were included in the analysis based on 224,295 person-months of follow-up. We used pooled logistic regression to assess the relationship between risk factors and rates of ON events. After identifying a set of variables related to ON incidence, we fit a final multivariable model to identify the most important risk factors for incident ON. RESULTS: In 18,691 person-years of follow-up after cohort entry, 122 incident ON events were observed (rate = 6.5/1,000 person-years). In the multivariable analysis, African American patients were at twice the risk for ON compared to White patients. Male patients and smokers had an increased risk for ON of ~80% and 50% compared to female patients and nonsmokers, respectively. For every 10-year increase in the age at diagnosis, there was a 20% reduced risk for ON. Patients diagnosed after the 1990s had a 50% reduced risk of ON compared to those diagnosed before the 1990s. A highest daily dosage of prednisone of 40 mg or higher, even when administered for a month or less, significantly increased the risk of ON. Use of pulse methylprednisolone or intramuscular triamcinolone was not associated with an increased risk of ON. CONCLUSION: African American patients with systemic lupus erythematosus are at double the risk of experiencing ON compared to White patients. Oral prednisone at 20-39 mg for more than 1 month, or 40 mg daily for even 1 month, at any point in the disease course, remained the most important glucocorticoid predictor of ON.