WHAT IS THE BEST SURGICAL OPTION FOR PERIPROSTHETIC JOINT INFECTION (PJI) DUE TO <i>CANDIDA</i> SPECIES?

Aim Candida species are uncommon pathogens causing prosthetic joint infection (PJI). This study evaluated the surgical management and outcome of Candida PJI. Methods Patients with EBJIS Definition confirmed PJI, due to Candida species, from 19 medical centres were assessed. Demographic, diagnostic, medical and surgical treatment and outcome data were collected. Results 269 patients were recruited with follow-up for at least one year. Mean age was 70.2 years (+/- 12.4) with 10.8% being immunocompromised. The most common fungal species were C. albicans (55.8%), C. parapsilosis (29.4%), C. glabrata (7.8%) and C. tropicalis (5.6%). Co-infection with bacteria occurred in 138 (51.3%) cases. DAIR was performed in 96 (36.2%) cases, with 169 (63.8%) having implant exchange or removal (76 one-stage, 78 two-stage, 11 removal/Girdlestone arthroplasty, 2 amputation). Patient demographics and antifungal therapy were similar in all surgical groups. Overall, treatment was successful in 156 (58%) cases. Failure was more likely in older patients (&gt;70 years; p=0.008) and those who had DAIR (OR 1.945; 1.156-3.279; p=0.004). Failure was less likely with C. parapsilosis infection compared to C. albicans (31.6% vs 48%; p=0.037). DAIR patients had more co-infection with bacteria (63.5% vs 47.4%; p=0.013) and more previous surgeries (median 4 vs 3; p=0.007), but multivariate analysis showed that these were not independent risk factors for failure. There was no difference in mortality between DAIR patients and those with other surgery (13.5% vs 17.7%; p=0.372). DAIR was successful in 45/96 (46.9%) cases compared to 110/169 (65.1%) cases with other surgery (p&lt;0.004). Early DAIR (surgery performed &lt;I month from implantation/infection onset) was not more effective than late DAIR (surgery performed after 1 month)(early DAIR 44.4% cure vs 63.9% cure in late DAIR; p=0.004). Two-stage revision was successful in 54/78 (69.2%), which was significantly better than DAIR (p=0.003). One-stage revision was successful in 51/76 (67.1%) patients; also significantly better than DAIR (p=0.002), but equivalent to two-stage revision (p=0.777). Conclusion DAIR was successful in less than half of patients with Candida PJI. We could not identify any subgroup which might have better outcomes with this surgical option. Interestingly, almost 90% of our patients with Candida PJI had no immunocompromise. One or two-stage revision offer a better option, if possible, and do not increase mortality.