Ruling Out Infection in a Painful Joint Replacement
A missed periprosthetic joint infection gets treated as a loose implant, and a new prosthesis placed into an infected joint is likely to fail again.
Key terminology
- Periprosthetic joint infection (PJI)
- This infection involves a joint replacement and the tissue around it. It can look exactly like a loose, painful implant with no infection.
- Aseptic loosening
- The implant comes loose from bone without any infection. It is the main look-alike the workup must separate from PJI.
- Sinus tract
- This draining channel runs from the skin down to the joint or implant. It confirms PJI on its own.
- Polymorphonuclear percentage (PMN%)
- This value is the share of white cells in the joint fluid that are neutrophils. A high value points toward bacterial infection.
- Alpha-defensin
- This peptide is released by neutrophils into joint fluid. A positive lab assay is highly specific for PJI.
- ICM 2018
- This 2018 International Consensus Meeting definition gives points for each abnormal test. A total of 6 or more means infected.
- EBJIS
- This 2021 European Bone and Joint Infection Society definition uses no points. It sorts patients into 3 tiers of unlikely, likely and confirmed.
Takeaways
- Every painful or failing joint replacement gets an infection workup before any revision. Pain and stiffness are the most sensitive signs of PJI and overlap with aseptic loosening. Fever and erythema are specific and insensitive. Early loosening is more often a hidden infection than late loosening (McNally et al, Bone Joint J 2021;103-B:18-25, PMID 33380199).
- Serum CRP and ESR are the first screen, and normal values do not exclude infection. ICM 2018 gives 2 points for CRP >1 mg/dL and 1 point for ESR >30 mm/h. Among 952 surgically treated PJIs, about 4% had a normal ESR and CRP, most often with coagulase-negative staph. Aspiration made the diagnosis in most of them (Parvizi et al, J Arthroplasty 2018;33:1309-1314.e2, PMID 29551303; McArthur et al, Bone Joint J 2015;97-B:939-944, PMID 26130349).
- Aspiration with synovial cell count and differential is the core test for chronic infection. ICM 2018 scores synovial WBC >3,000 cells/µL and polymorphonuclear percentage (PMN%) >80%, and EBJIS uses the same cutoffs to confirm PJI in the hip and knee. Published knee cutoffs run lower (>1,100 cells and >64%). A preoperative culture is insensitive and cannot rule PJI out (Parvizi et al, J Arthroplasty 2018;33:1309-1314.e2, PMID 29551303; McNally et al, Bone Joint J 2021;103-B:18-25, PMID 33380199; Ghanem et al, J Bone Joint Surg Am 2008;90:1637-1643, PMID 18676892).
- A sinus tract or 2 cultures growing the same pathogen confirms infection. ICM 2018 lists both as major criteria. EBJIS calls a sinus tract communicating with the joint (or an exposed implant) the only fully specific clinical finding. A single culture of a skin contaminant such as coagulase-negative staph does not confirm PJI on its own (Parvizi et al, J Arthroplasty 2018;33:1309-1314.e2, PMID 29551303; McNally et al, Bone Joint J 2021;103-B:18-25, PMID 33380199).
- Antibiotics given before aspiration can hide the pathogen. Prior antibiotic use was associated with higher odds of culture-negative PJI (OR 4.7). EBJIS asks that patients be off antibiotics at least 2 weeks before intraoperative sampling, and prophylaxis at incision is still given. Alpha-defensin may help when antibiotics were already started (Malekzadeh et al, Clin Orthop Relat Res 2010;468:2039-2045, PMID 20401555; McNally et al, Bone Joint J 2021;103-B:18-25, PMID 33380199).
- Chronic cutoffs overcall infection in the first weeks after surgery. Hemarthrosis and postoperative inflammation raise the cell count. Within 6 weeks of primary TKA, the best synovial WBC cutoff was 27,800 cells/µL (PMN 89%, CRP 95 mg/L). Using the usual 3,000 cutoff would have led to unnecessary reoperations (Bedair et al, Clin Orthop Relat Res 2011;469:34-40, PMID 20585914).
- ICM 2018 adds weighted points, and EBJIS sorts patients into tiers. Under ICM 2018 a score of 6 or more is infected and a combined score of 4-5 is inconclusive. EBJIS uses infection unlikely, likely and confirmed, calls scores arbitrary and drops ESR and D-dimer as adding nothing over CRP. ICM 2018 had the support of only 68% of delegates and was not endorsed by MSIS or EBJIS (Parvizi et al, J Arthroplasty 2018;33:1309-1314.e2, PMID 29551303; McNally et al, Bone Joint J 2021;103-B:18-25, PMID 33380199).
- EBJIS labels more joints infected and leaves fewer in the gray zone. In 472 revisions, PJI was diagnosed in 41% by EBJIS, 40% by IDSA, 36% by ICM 2018 and 31% by MSIS. In 206 revisions, EBJIS caught every PJI found by the other definitions and left 8% likely vs 15% inconclusive by ICM. Both studies included EBJIS authors (Sousa et al, Acta Orthop 2023;94:8-18, PMID 36701120; Sigmund et al, Bone Joint Res 2022;11:608-618, PMID 36047011).