2005 Abstract : Cl 1

Authors: Sangita Dash MD, Juanmanuel Gomez MD, J. Robert Cantey MD, Camelia Marculescu MD

Title: OUTCOMES AND RISK FACTORS FOR FAILURE OF PROSTHETIC JOINT INFECTIONS (PJI) AT A TERTIARY CARE CENTER

Addresses: 100 Doughty Street, Suite 210 BA/IOP South, Charleston, SC 29425

Purpose: To determine the outcome and associated risk factors for treatment failure of PJI episodes at a tertiary care center having different demographic characteristics than previously reported cohorts.

Methods: All episodes of total hip arthroplasty (THA) and total knee arthroplasty (TKA) PJI diagnosed using strict case definition between 1/98 and 12/03 were analyzed. Failure was defined as reinfection with a different strain or different microorganism, relapse of infection, superinfection, presence of a sinus tract, purulence in a joint space, acute inflammation on a pathological exam of periprosthetic tissue and indeterminate clinical failure. Variables considered to be risk factors for failure were entered in a multivariable logistic regression model in a stepwise fashion.

Results: There were 117 episodes of PJI in 103 patients. The median age of this cohort was 66 (range, 28-86). Thirty-eight (38%) percent of the episodes occurred in males and 33% occurred in African Americans. Fifty-six of 117 (48%) involved a THA. Fifty-five of 117 (47%) of episodes involved a primary joint arthroplasty. Rheumatoid arthritis, diabetes mellitus, and systemic malignancy were present in 11%, 28%, and 9% of the episodes, respectively. Pain, swelling, and sinus tract were present in 78.6%, 47%, and 22% of all PJI episodes respectively. Median duration of symptoms was 19 days (range, 1-1500). Median joint age was 436 days (range, 10-10132). Of the 117 episodes of PJI, 51 (44%) had retained hardware. Of these, 47% (24/51) failed. Eighteen of 66 (27%) episodes that had hardware removed failed. Thirty-one of 117 (26%) of the PJI episodes were due to Staphylococcus aureus (SA), 28 of 117 (24%) were due to coagulase negative staphylococci, 13 of 117 (11%) were due to streptococci, and 22 of 117 (19%) were polymicrobial. Ten of 117 (8.5%) of all PJI episodes were due to methicillin-resistant SA (MRSA), including MRSA in polymicrobial infections. The median duration of effective antimicrobial therapy was 42 days (range, 9-175). Vancomycin and B-lactams were the main intravenous therapy, used in 64/117 (55%) and 44/117 (38%) of all episodes. Chronic oral suppression was used only in 7 of 117 (6%) of all episodes. In multivariable logistic regression model, retained hardware (OR = 4.1, 95%CI:1.69-10.86), infection of a TKA (OR = 4.0, 95%CI:1.63-10.77), and MRSA PJI
(OR = 4.4, 95%CI:1.06-20.93) were risk factors significantly associated with treatment failure.

Conclusion: In our data, PJI due to MRSA, presence of retained hardware, and infection of TKA were significant risk factors for treatment failure.

Significance: This paper provides evidence that infection with MRSA is a risk factor for treatment failure in episodes of PJI involving both primary and revision arthroplasties, by using different definitions of failure than previously reported. Prospective studies are needed to confirm our finding.