Accuracy of the International Consensus Scoring System for diagnosis of periprosthetic joint infection in revision shoulder arthroplasty

Abstract

BackgroundTotal shoulder arthroplasty (TSA) continues to undergo dramatic growth with expanding indications and improvements in implants and surgical techniques. A major complication following TSA is periprosthetic joint infection (PJI), which remains difficult to diagnose, often relying on clinical judgment. A contemporary definition of PJI was established at the 2018 International Consensus Meeting (ICM) on Musculoskeletal Infection. We sought to retrospectively examine the accuracy of this scoring system in previously performed revision TSA and hypothesized that the ICM scoring system would be reliable in determining the presence of TSA PJI. MethodsOur institutional database was reviewed to identify patients undergoing revision TSA before the advent of the ICM PJI scoring system. Clinical notes and operative reports were reviewed for data regarding the preoperative clinical examination, laboratory values, and intraoperative findings. The findings were assigned scores based on the definition of probable PJI by the ICM scoring system. Scores were compared to treatment plans of infected vs. noninfected patients. The diagnosis of PJI was made using a combination of clinical examination, laboratory values, and intraoperative findings. Sensitivity, specificity, positive and negative predictive values, and accuracy of the ICM scoring system were calculated compared to actual treatment decision, the gold standard. ResultsOf 81 revision arthroplasties, 52 were revision reverse TSA (rTSA), and 29 were revision anatomic TSA (aTSA). Seven rTSA patients were treated as infected (7/52, 13.5%), and the scoring system identified 4 of those as being probable infections (4/7, 57.1%). One additional rTSA patient scored as probable infection, underwent a revision for instability, and was found to have no infection. Three aTSA patients were treated as infected (3/29, 10.3%), with one of those identified as probable infection by the scoring system (1/3, 33.3%). Four patients in the rTSA group and no patients in the aTSA group met the criteria for definite infection. Using the threshold of probable infection to identify PJI, the sensitivity of the scoring system was 0.6, and specificity was 0.99. The positive predictive value was 0.86, and the negative predictive value was 0.95. With the same threshold, the ICM scoring system was 93.8% accurate. ConclusionsIdentifying PJI in TSA remains difficult in the absence of definite signs of joint sepsis. This study found the scoring system to be highly accurate, although with modest sensitivity, and a reliable tool for the diagnosis of PJI following TSA. Level of evidenceLevel IV; Retrospective Case Series with No Comparison Group Treatment Study