Association of tramadol with all-cause mortality, cardiovascular diseases, venous thromboembolism, and hip fractures among patients with osteoarthritis: a population-based study
Lingyi Li,Shelby Marozoff,Na Lu,Hui Xie,Jacek A Kopec,Jolanda Cibere,John M Esdaile,J Antonio Aviña-Zubieta
DOI: https://doi.org/10.1186/s13075-022-02764-3
2022-04-11
Abstract:Background: The use of tramadol among osteoarthritis (OA) patients has been increasing rapidly around the world, but population-based studies on its safety profile among OA patients are scarce. We sought to determine if tramadol use in OA patients is associated with increased risks of all-cause mortality, cardiovascular diseases (CVD), venous thromboembolism (VTE), and hip fractures compared with commonly prescribed nonsteroidal anti-inflammatory drugs (NSAIDs) or codeine. Methods: Using administrative health datasets from British Columbia, Canada, we conducted a sequential propensity score-matched cohort study among all OA patients between 2005 and 2013. The tramadol cohort (i.e., tramadol initiation) was matched with four comparator cohorts (i.e., initiation of naproxen, diclofenac, cyclooxygenase-2 [Cox-2] inhibitors, or codeine). Outcomes are all-cause mortality, first-ever CVD, VTE, and hip fractures within the year after the treatment initiation. Patients were followed until they either experienced an event, left the province, or the 1-year follow-up period ended, whichever occurred first. Cox proportional hazard models were used to estimate hazard ratios after adjusting for competing risk of death. Results: Overall, 100,358 OA patients were included (mean age: 68 years, 63% females). All-cause mortality was higher for tramadol compared to NSAIDs with rate differences (RDs/1000 person-years, 95% CI) ranging from 3.3 (0.0-6.7) to 8.1 (4.9-11.4) and hazard ratios (HRs, 95% CI) ranging from 1.2 (1.0-1.4) to 1.5 (1.3-1.8). For CVD, no differences were observed between tramadol and NSAIDs. Tramadol had a higher risk of VTE compared to diclofenac, with RD/1000 person-years (95% CI) of 2.2 (0.7-3.7) and HR (95% CI) of 1.7 (1.3-2.2). Tramadol also had a higher risk of hip fractures compared to diclofenac and Cox-2 inhibitors with RDs/1000 person-years (95% CI) of 1.9 (0.4-3.4) and 1.7 (0.2-3.3), respectively, and HRs (95% CI) of 1.6 (1.2-2.0) and 1.4 (1.1-1.9), respectively. No differences were observed between tramadol and NSAIDs for all events. Conclusions: OA patients initiating tramadol have an increased risk of mortality, VTE, and hip fractures within 1 year compared with commonly prescribed NSAIDs, but not with codeine.