Periprosthetic joint infection (PJI): surgeons have built algorithms, convened panels, and memorized criteria from the International Consensus Meeting. And yet, one question lingers in the clinic:
Can you really trust synovial C-reactive protein test (CRP)?
A new study takes a hard look at the performance of synovial C-reactive protein in diagnosing PJI after total knee arthroplasty (TKA) and total hip arthroplasty (THA). The verdict? Let’s just say synovial CRP may not be the diagnostic superhero you hoped for.
Testing the Tester
The authors analyzed 107 patients (86 TKAs and 21 THAs), testing for possible PJI. Using the 2018 International Consensus Meeting criteria as the diagnostic gold standard, they examined how well synovial CRP performed at the commonly cited cutoff of >6.9mg/L.
They calculated the usual suspects: Sensitivity, specificity and accuracy.
Knees: Specific but Not Sensitive
In total knee arthroplasty: Sensitivity: 68.8%, specificity: 97.8% and accuracy: 85.9%.
If synovial CRP is high, you can be pretty confident you’ve got infection. But if it’s low? Don’t pop the champagne just yet.
Nearly one-third of infected knees could be missed at that threshold.
Hips: A Similar Story
In total hip arthroplasty: Sensitivity: 75.0%, specificity: 92.3% and accuracy: 85.7%.
Again, solid specificity. But sensitivity still leaves room for uncomfortable false negatives.
Lower the Threshold?
Using Youden’s J-statistic, the authors identified new “optimal” cutoffs: Knees: 3.43mg/L and Hips: 2.86mg/L
Did lowering the bar improve sensitivity? Well…yes and no.
Because when synovial CRP went head-to-head with serum CRP in ROC (Receiver Operating Characteristic) analysis, serum CRP actually held its own — and in knees, even edged ahead.
Knees Area Under the Curve: Synovial CRP: 0.910, serum CRP: 0.954.
Hips Area Under the Curve: Synovial CRP: 0.885, serum CRP: 0.886.
And just to make things more interesting, synovial and serum CRP were strongly correlated in both joints (ρ ≈ 0.80).
Which raises an uncomfortable but practical question: If serum CRP performs as well — or better — and correlates strongly with synovial CRP…are you gaining much by tapping the joint just to measure CRP?
The Clinical Takeaway: Caution, Not Blind Faith
The traditional 6.9mg/L synovial CRP cutoff appears too insensitive, risking false reassurance in infected joints. Lower thresholds improve sensitivity but complicate interpretation and may affect specificity.
Meanwhile, serum CRP remains a strong and, frankly, easier player in the diagnostic game.
Bottom Line
- A synovial CRP >6.9mg/L is convincing — but not sensitive enough.
- Lowering the cutoff improves detection but changes the balance.
- Serum CRP performs at least as well and correlates strongly.
- Synovial CRP should be interpreted as one data point, not a verdict.
In the high-stakes chess match of PJI diagnosis, synovial CRP is a useful piece — but it’s not the queen.
And as always in arthroplasty: trust the whole board, not just one move.
Origin Study Title: The Limited Utility of Synovial C-Reactive Protein in the Diagnosis of Knee and Hip Periprosthetic Joint Infection: Proposed New Diagnostic Cutoffs
Authors:
David McCavitt, B.A.; McKenzie W. Culler, B.A.;∙Ryan C. Palmer, M.D.; Avinash Iyer, B.S.; Matthew A. Lim, B.S.; Jay Lieberman, M.D.; Nathanael D. Heckmann, M.D.
