Non-steroidal anti-inflammatory drugs and the risk of a second hip fracture: a propensity-score matching study

Po-Yao Chuang, Shih-Hsun Shen, Tien-Yu Yang, Tsan-Wen Huang, Kuo-Chin Huang

Non-steroidal anti-inflammatory drugs (NSAIDs) are frequently prescribed for elderly patients, particularly after a hip fracture. However, we are not clear about the effect of NSAIDs on the risk of a second hip fracture because of confounding factors. This was a Taiwan National Health Insurance Research Database-based study using propensity-score matching (PSM) to control for confounding. Enrollees were selected from patients with a hip fracture during 1996-2004 and followed longitudinally until December 2009. After PSM for comorbidities and bisphosphonate therapy, 94 patients with a second hip fracture were assigned to the Cases group and 461 without it to the Controls group. The target drugs are NSAIDs; paracetamol and dexamethasone are used for comparison. The correlation between the mean daily-dose (MDD) ratios of NSAIDs and the probability values of the current statistical tests were highly negative (Pearson's r = -0.920, P = 0.003), which indicated that the higher the MDD ratios, the greater the risks of a second hip fracture. A Kaplan-Meier survival analysis showed a time-dependent trend of increasing risk of a second hip fracture in patients taking NSAIDs (P < 0.001). Moreover, patients ≥60 years old had a higher risk of a second hip fracture than did those <60 and taking the NSAIDs diclofenac (P = 0.016) and celecoxib (P = 0.003) and the corticosteroid dexamethasone (P = 0.018), but not those taking analgesic paracetamol (P = 0.074). We conclude that taking NSAIDs after a fragility hip fracture dose- and time-dependently significantly increases the risk of a second hip fracture, especially in elderly patients. To lower the risk of a second hip fracture, any underlying causes for excessively using NSAIDs should be treated and thus fewer NSAIDs prescribed after a first hip fracture.