The Journal of Arthroplasty, Volume 35, Issue 3, 849 - 854

Treatment Outcomes and Attrition in Gram-Negative Periprosthetic Joint Infection

Kalbian, Irene L. et al.
Hip Knee


While the prevailing belief is that periprosthetic joint infection (PJI) caused by Gram-negative (GN) organisms confers a poorer prognosis than Gram-positive (GP) cases, the current literature is sparse and inconsistent. The purpose of this study is to compare the treatment outcomes for GN PJI vs GP PJI and Gram-mixed (GM) PJI.


A retrospective review of 1189 PJI cases between 2007 and 2017 was performed using our institutional PJI database. Treatment failure defined by international consensus criteria was compared between PJI caused by GN organisms (n = 45), GP organisms (n = 663), and GM (n = 28) cases. Multivariate regression was used to predict time to failure.


GM status, but not GN, had significantly higher rates of treatment failure compared to GP PJI (67.9% vs 33.2% failure; hazards ratio [HR] = 2.243, P = .004) in the multivariate analysis. In a subanalysis of only the 2-stage exchange procedures, both GN and GM cases were significantly less likely to reach reimplantation than GP cases (HR = .344, P < .0001; HR = .404, P = .013).


Although there was no observed difference in the overall international consensus failure rates between GN (31.1% failure) and GP (33.2%) PJI cases, there was significant attrition in the 2-stage exchange GN cohort, and these patients were significantly less likely to reach reimplantation. Our findings corroborate the prevailing notion that GN PJI is associated with poorer overall outcomes vs GP PJI. These data add to the current body of literature, which may currently underestimate the overall failure rates of GN PJI treated via 2-stage exchange and fail to identify pre-reimplantation morbidity.

Download article