Outcomes following Revision Total Knee Arthroplasty Septic versus Aseptic Failure: A National Propensity-Score-Matched Comparison.

Abstract

Although periprosthetic joint infection (PJI) has been hypothesized to increase the risk of complications following revision total knee arthroplasty (TKA), strong evidence linking the two is lacking. The aim of this study was to determine whether PJI is an independent risk factor for inpatient perioperative complications, and increased resource use in patients undergoing revision TKA. We relied on the US Nationwide Inpatient Sample (NIS) to identify patients with PJI or non-PJI treated with revision TKA between 2002 and 2014. Overall, 5,316 (16.4%) and 27,033 (83.6%) patients were categorized as PJI and non-PJI, respectively. To adjust for potential baseline differences between the two groups, propensity-score-based matching was performed. This resulted in 5,187 (50%) PJI patients matched to 5,187 (50%) non-PJI patients. The rates of postoperative complications, blood transfusions, prolonged length of stay (pLOS), in-hospital cost, and in-hospital mortality were assessed for both groups. Multivariable logistic regression analyses were performed within the cohort after propensity-score matching. For PJI versus non-PJI, respectively, the following rates were recorded: blood transfusions, 28.3 versus 18.4% (<i>p</i> < 0.0001); postoperative complications, 27.5 versus 19.8% (<i>p</i> < 0.0001); pLOS (>4 d), 61.9 versus 26.9% (<i>p</i> < 0.0001); in-hospital cost (>$39,929), 55.8 versus 44.2% (<i>p</i> < 0.0001); in-hospital mortality, 0.6 versus 0.3% (<i>p</i> = 0.016). In multivariable logistic regression analyses, PJI patients were more likely to receive a blood transfusion (odds ratio [OR]: 1.78; <i>p</i> < 0.0001), to experience postoperative complications (OR: 1.56; <i>p</i> < 0.0001), to have a higher in-hospital cost (OR: 1.65; <i>p</i> < 0.0001), to have a pLOS following surgery following surgery (OR: 4.69; <i>p</i> < 0.0001), and to have a higher in-hospital mortality (OR: 2.14; <i>p</i> = 0.019). After adjustment for potential selection biases, PJI is associated with more adverse perioperative outcomes and resource use than non-PJI patients. This is a Level II (level of evidence), prognostic study.