Risk of Venous Thromboembolism After Total Knee Arthroplasty in Patients with Obstructive Sleep Apnea: Results from a National Cohort
Cosmo Fowler,Simar Chawla,Lauren Chism,Stephen M Pastores,Dennis H Auckley
DOI: https://doi.org/10.1213/ANE.0000000000007167
2024-09-16
Abstract:Background: Obstructive sleep apnea (OSA) is a prevalent condition associated with many comorbidities. However, establishing the independent impact of OSA on specific health outcomes can be challenging without access to a substantial patient cohort. This study aimed to investigate whether a diagnosis of OSA was independently associated with venous thromboembolism (VTE) after total knee arthroplasty (TKA). Methods: In this retrospective cohort study, we interrogated the TriNetX Analytics Research Network, a large database comprising the billing claims and electronic health record-derived data of >117 million patients. Using Current Procedural Terminology (CPT) and International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes, we identified US adult patients who underwent TKA between January 1, 2013 and January 1, 2023, with and without preexisting OSA (and ≥2 OSA occurrences overall). We then analyzed the 1-month postoperative incidence of VTE as a composite outcome of deep vein thrombosis (DVT) and pulmonary embolism (PE) incidence, as well as cerebrovascular accident (CVA), myocardial infarction (MI), and DVT and PE individually. Baseline demographic and comorbidity covariates were incorporated into a 1:1 propensity score-matched analysis to clarify the independent effect of OSA. Results: During the 10-year study period, a total of 197,460 patients underwent TKA. Of these, 27,976 met the criteria for inclusion in the OSA cohort, while 150,830 had no documented history of OSA. In the initial analysis, OSA was significantly associated with the primary outcome (DVT/PE) as well as all secondary outcomes (CVA, MI, and individually with DVT and PE) at 1 month postsurgery. After generating propensity score matched cohorts, DVT/PE remained significantly associated with OSA, with an absolute risk difference of 0.7% (odds ratio [OR], 1.19, confidence interval [CI], 1.1-1.3, P < .001), as were the secondary outcomes of DVT (OR, 1.11, CI, 1.0-1.2, P = .030) and PE (OR, 1.41, CI, 1.2-1.6, P < .001). Conclusions: In this study encompassing a nationally representative sample of TKA patients, OSA was associated with increased incidence of VTE at 1 month postoperatively, an association that persisted after the generation of matched cohorts. While limitations related to the lack of patient-level data, disease severity, and therapy adherence should be acknowledged, our large sample size enabled us to factor many baseline characteristics into our analysis, reinforcing the association of these findings. Prospective work is needed on the impact of modulating factors such as anticoagulation regimen and positive airway pressure therapy on these outcomes.