Abstract / Summary
Abstract Objective The safety of lumbar puncture in patients with a history of anticoagulant or antiplatelet use remains incompletely characterized. This study evaluated the coded occurrence of blood patch, paralysis, and spinal decompression within 30 days of lumbar puncture using a retrospective database. Methods TriNetX, a database that houses clinical information from multiple healthcare entities, was used to carry out retrospective data collection and analysis. Data from patients with a history of antiplatelet use and anticoagulant use were collected separately. The timeline for outcomes of blood patch, paralysis, and spinal decompression started on the day of the first lumbar puncture and ended on the 30th day after the procedure. Propensity score matching was followed by statistical risk analyses. Results A total of 13,310 patients were identified after propensity score matching for both lumbar puncture patients with antiplatelet use and lumbar puncture patients without antiplatelet, anticoagulant, heparin, or low-molecular-weight heparin use. Blood patch (OR [95% CI], 0.705 [0.563–0.882]) showed lower odds of association, while paralysis (OR [95% CI], 1.404 [1.097–1.798]) and spinal decompression (OR [95% CI], 2.225 [1.183–4.185]) showed greater odds of association; all were statistically significant. The corresponding absolute risk differences were 0.3% for paralysis and 0.1% for spinal decompression. A total of 25,788 patients were identified for both lumbar puncture patients with anticoagulant use and those without blood-thinner use after propensity score matching. Blood patch (OR [95% CI], 0.394 [0.332–0.467]) and spinal decompression (OR [95% CI], 0.659 [0.389–1.115]) showed lower odds of association, while paralysis (OR [95% CI], 1.154 [0.943–1.414]) showed greater odds of association; only blood patch was statistically significant. Conclusions Among patients with a history of antiplatelet use, coded paralysis and spinal decompression within 30 days of lumbar puncture were associated with higher odds, although the absolute risk differences were small at 0.3% and 0.1%, respectively. Anticoagulant use history was not significantly associated with either paralysis or spinal decompression. Blood patch use was significantly lower in both exposure analyses. Since the TriNetX database cannot establish whether these coded outcomes were caused by lumbar puncture, these associations should be interpreted cautiously and require chart-level confirmation.