Abstract
Aim
Knee arthrodesis (KA) and above knee amputation (AKA) have been used for salvage of failed total knee arthroplasty (TKA) in the setting of periprosthetic joint infection (PJI). The factors that lead to a failed fusion and progression to AKA are not well understood. The purpose of our study was to determine factors associated with failure of a staged fusion for PJI and predictive of progression to AKA.
Method
We retrospectively reviewed a single-surgeon series of failed TKA for PJI treated with two-stage KA between 2000 and 2016 with minimum 2-year follow-up. Patient demographics, comorbidities, surgical history, tissue compromise, and radiographic data were recorded. Outcomes were additional surgery, delayed union, Visual Analog Pain scale (VAS) and Western Ontario and McMaster Activity score (WOMAC). No power analysis was performed for this retrospective study. Medians are reported as data were not normally distributed.
Results
Fifty-one knees underwent fusion with median follow-up of 7 years (interquartile range (IQR) of 2–18 years). Median age was 71 years old (IQR 47 – 98), with a M:F ratio of 23:28. Median BMI was 34.3 kg/m2 (IQR 17.9–61). Infection was eradicated in 47 knees (92.2%); 24 knees (47.0%) required no additional surgery. 41 patients (83.6%) remained ambulatory after knee fusion, with 21% of these patients (10 total) requiring no ambulatory assistive device. Median VAS following arthrodesis was 4.6 (range 0–10). Median WOMAC was 36.2 (range 9–86). Three TKAs (5.9%) underwent AKA for overwhelming infection. Predictors of AKA were chronic kidney disease (OR 4.0, 95% CI 0.6–26.8), peripheral vascular disease (OR 3.5, 95% CI 0.3–44.7), AORI III bone loss (OR 2.6, 95% CI 0.4–35.2), instability (OR 2.2, 95% CI 0.2–15.9), and immunosuppression (OR 1.1, 95% CI 0.1–7.8). Tobacco use (OR 8.6, 95%CI 2.4–31.4), BMI>25 (OR 3.8, 95% CI 0.43–32.5) and instability prior to arthrodesis (OR 2.51, 95% CI 0.77–8.21) were associated with non-union. All other risk factors (gender, diabetes mellitus, chronic kidney disease, peripheral vascular disease, massive bone stock loss, and immunosuppression) were not associated with arthrodesis failure.
Conclusions
Staged KA for PJI in severely compromised hosts provides a functional limb free of infection and rarely results in conversion to AKA. Given our small sample size, ability to establish statistical significance of predictive factors for AKA after PJI was limited, but CKD, peripheral vascular disease, AORI III bone loss, instability, and immunosuppression trended towards significance as predictors of failure of KA after PJI predisposing to AKA.
