Association of Perioperative Dexamethasone With Postoperative Complications After Primary Total Joint Arthroplasty: An Instrumental Variable Analysis

BACKGROUND: Retrospective studies suggest that dexamethasone may provide benefits that extend beyond its antiemetic properties, including a reduction in postoperative complications. However, results from randomized controlled trials have not consistently shown there to be a reduction in composite major adverse events. This discrepancy may be due to confounding factors, measurement error, or simultaneity bias among retrospective investigations. This study used instrumental variable analysis (IVA) to help address potential sources of bias and better estimate treatment effects in patients undergoing total joint arthroplasty (TJA). METHODS: Patients who underwent primary elective TJA between 2016 and 2021 were identified using diagnosis and procedural codes. Bivariate regression, multivariable regression, and IVA were conducted. The primary end point was a 90-day composite (any versus none) of major postoperative medical complications. Secondary outcomes were infection, readmission, and death. Two distinct instruments—the frequency of dexamethasone use by surgeon and by hospital—were used to evaluate the robustness of our IVA. Patient demographics, hospital factors, and comorbidities were reported using descriptive statistics. Instrumental variable covariates were selected using the least absolute shrinkage and selection operator with 3 regularization parameter strategies. RESULTS: 1525,844 TJAs performed between 2015 and 2021 were identified (976,996 knees [total knee arthroplasty {TKA}]; 548,848 hips [total hip arthroplasty {THA}]). Major postoperative medical complications were observed in 31,299 (3.43%) dexamethasone-exposed patients compared to 31,266 (4.87%) unexposed patients. Surgeon-based IVA yielded results comparable to the multivariable and bivariate analysis (local average treatment effect [LATE]: TKA: –1.20% [95% confidence interval [CI], –1.33% to –1.08%]; THA: –1.14% [95% CI, –1.30% to –0.99%]). Hospital-based IVA produced similar findings (LATE: TKA: –1.23% [95% CI, –1.38% to –1.09%]; THA: –1.18% [95% CI, –1.35% to –1.00%]). Both instruments demonstrated high F -statistics and significant Hausman tests. Secondary outcomes mirrored these results, except for mortality, which did not meet endogeneity criteria across analyses. CONCLUSIONS: The findings of this study support that dexamethasone exposure is associated with a reduction in composite major postoperative complications after TJA. The observed moderate treatment effect, in conjunction with a low baseline incidence of adverse events, may explain the inconsistent outcomes reported in previous randomized trials. Future prospective studies should incorporate composite end points and target high-risk patient populations or procedural subgroups.