26Postoperative diagnosis and outcome in patients with revision arthroplasty for aseptic loosening

Marta Fernandez-Sampedro, Carlos Salas-Venero, Concepción Fariñas-Álvarez, Manuel Sumillera, Luis Pérez-Carro, Michel Fakkas-Fernandez, Javier Gómez-Román, Luis Martínez-Martínez, María Carmen Fariñas

The most common cause of implant failure is aseptic loosening (AL), followed by prosthetic joint infection (PJI). This study evaluates the incidence of PJI among patients operated with suspected AL and whether the diagnosis of PJI was predictive of subsequent implant failure including re-infection, at 2 years of follow up. Patients undergoing revision hip or knee arthroplasty due to presumed AL from February 2009 to September 2011 were prospectively evaluated. A sonication fluid of prosthesis and tissue samples for microbiology and histopathology at the time of the surgery were collected. Implant failure include recurrent or persistent infection, reoperation for any reason or need for chronic antibiotic suppression. Of 198 patients with pre-and intraoperative diagnosis of AL, 24 (12.1 %) had postoperative diagnosis of PJI. After a follow up of 31 months (IQR: 21 to 38 months), 9 (37.5 %) of 24 patients in the PJI group had implant failure compared to only 1 (1.1 %) in the 198 of AL group (p < 0.0001). Sensitivity of sonicate fluid culture (>20 CFU) and peri-prosthetic tissue culture were 87.5 % vs 66.7 %, respectively. Specificities were 100 % for both techniques (95 % CI, 97.9-100 %). A greater number of patients with PJI (79.1 %) had previous partial arthroplasty revisions than those patients in the AL group (56.9 %) (p = 0.04). In addition, 5 (55.5 %) patients with PJI and implant failure had more revision arthroplasties during the first year after the last implant placement than those patients with PJI without implant failure (1 patient; 6.7 %) (RR 3.8; 95 % CI 1.4-10.1; p = 0.015). On the other hand, 6 (25 %) patients finally diagnosed of PJI were initially diagnosed of AL in the first year after primary arthroplasty, whereas it was only 16 (9.2 %) patients in the group of true AL (RR 2.7; 95 % CI 1.2-6.1; p = 0.03). More than one tenth of patients with suspected AL are misdiagnosed PJI. Positive histology and positive peri-implant tissue and sonicate fluid cultures are highly predictive of implant failure in patients with PJI. Patients with greater number of partial hip revisions for a presumed AL had more risk of PJI. Early loosening is more often caused by hidden PJI than late loosening.