Validation of the Incidence of Reported Periprosthetic Joint Infections in Total Hip and Knee Arthroplasty in the Dutch Arthroplasty Register

Abstract

BackgroundArthroplasty registers underreport the incidence of periprosthetic joint infections (PJIs). We validated the incidence of reported PJIs in total hip arthroplasties (THAs) and total knee arthroplasties (TKAs) in the Dutch Arthroplasty Register (LROI) using data from the Dutch National Nosocomial Surveillance Network (PREZIES). MethodsAll primary THAs and TKAs from the LROI and all primary THAs and TKAs performed in consenting hospitals from PREZIES between 2012 and 2018 were matched on date of birth, date of surgery, sex, hospital, and type of procedure (THA n = 91,208; TKA n = 80,304). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for PJIs registered in the LROI, using PREZIES as a reference. ResultsThe incidence of registered PJIs in THAs was 1.2% in PREZIES and 0.5% in the LROI. For TKAs, this was 0.7 and 0.4%, respectively. The PJIs in THAs in the LROI had a sensitivity of 0.32 (confidence interval [CI]: 0.29 to 0.35), specificity of 1.00 (CI: 1.00 to 1.00), PPV of 0.74 (CI: 0.70 to 0.78), and NPV of 0.99 (CI: 0.99 to 0.99). In TKAs, the sensitivity, specificity, PPV, and NPV were 0.38 (CI: 0.34 to 0.42), 1.00 (CI: 1.00 to 1.00), 0.65 (CI: 0.59 to 0.70), and 1.00 (CI: 1.00 to 1.00), respectively. ConclusionsThe LROI captures approximately one-third of the PJIs as revision within one year for infection or resection arthroplasty. The capture rate of PJIs can be improved by including all reoperations without component exchange and nonsurgical treatments with antibiotics only.