Rapid reduction in fracture risk after the discontinuation of long-term oral glucocorticoid therapy: a retrospective cohort study using a nationwide health insurance claims database in Japan

The Summary Increased fracture risk due to oral glucocorticoids (GCs) rapidly decreases with GC discontinuation. However, evidence for this is limited. We found that fracture risk decreased rapidly in the first year after GC discontinuation, while hip fracture risk remained higher than reference levels for about two years after GC discontinuation. Purpose We investigated changes in fracture risk following discontinuation of long-term oral glucocorticoids (GCs) using Japan's nationwide health insurance claims database (NDBJ). Methods We identified patients aged >= 50 years who initiated GC therapy in 2012-2019. Those receiving >= 5 mg (prednisolone or equivalent, PSL)/day for >= 72 days in the initial 90 days of GC therapy were classified as the GC-exposure group, and those receiving < 5 mg PSL/day for < 30 days were classified as the reference group. Patients discontinuing GC after 90 days of GC therapy were classified as the GC-discontinuation group; all others were classified as the GC-continuation group. We tracked the incidence rates of hip and clinical vertebral fractures for up to 990 days, and assessed fracture risk after GC discontinuation by hazard ratios (HR) adjusted by inverse probability weighting using propensity scores for GC discontinuation. Results There was a total of 52,179 GC-discontinuation, 91,969 GC-continuation, and 43,138 reference group women, and 57,560, 93,736, and 33,696 men in the corresponding groups, respectively. According to adjusted HRs, incidence rates of fractures were significantly lower in the GC-discontinuation group than in the GC-continuation group in the initial 90 days after GC discontinuation and remained significant for 360 days, except for hip fracture in men. HRs for hip fractures remained significantly higher in the GC-discontinuation group compared to the reference group for 720 days post-discontinuation. Conclusion Fracture risk declines rapidly in the first year after GC discontinuation, but vigilance is necessary as the increased risk persists for two years post-discontinuation.