Infection: It's a Numbers Game: Commentary on an article by William Z. Stone, MD, et al.: "Clinical Evaluation of Synovial Alpha Defensin and Synovial C-Reactive Protein in the Diagnosis of Periprosthetic Joint Infection".

Abstract

Commentary Our ability to continue to safely perform arthroplasty depends on our ability to control infection. With bacterial resistance, we may need progress just to keep up. Progress depends on definitions, measures, and qualitative and quantitative clarity. So far, we do not agree on much about revision arthroplasty with the exception of the impact of periprosthetic joint infection1. The road to understanding is long, but in their article, Stone et al. advance the journey. Not all constructive ideas can be retrieved on an electronic search. Readers should correct me when I attribute the idea that “all arthroplasties are infected all the time” to Professor Reg Elson of the University of Sheffield (past president of the British Hip Society). With some mischief, he suggested that bacteria are always present in an arthroplasty, “in balance” with the patient’s immune surveillance. Upset the equilibrium, and periprosthetic joint infection results. Recent research into the human microbiome supports the complex relationship between humans and bacteria2. When does that relationship become a periprosthetic joint infection? Surgeons categorize everything as sterile or contaminated: there are either no bacteria or some bacteria. Sharper minds recognize that clinical infection is a numbers game. The classic laboratory study by Zimmerli et al. showed that rats tolerate a much larger inoculum without clinical infection in the absence of foreign material, but sustain infection with many fewer bacteria when metal implants are present3. It makes sense that periprosthetic joint infection can be diagnosed by the host’s response, rather than the bacteria themselves. C-reactive protein (CRP) appears in reaction to the C-polysaccharide of Pneumococcus and was identified decades ago4. We depend on this laboratory test daily when infection is suspected. Another marker, alpha defensin, can diagnose periprosthetic joint infection, but how dependable is it? The article by Stone et al. originates from clinicians who did not develop the alpha defensin assay. They evaluated the ability of synovial fluid alpha defensin alone and in combination with synovial fluid CRP to identify periprosthetic joint infection as defined by Musculoskeletal Infection Society (MSIS) criteria. The assays were good, but not a perfect match. The synovial fluid alpha defensin missed 7 infections, mainly with low-virulence organisms. When synovial fluid CRP was added, 3 of the 7 infected cases were still overlooked. Although patients were advised to stop antibiotics before the assay, there is no way to verify if they did or the effect of exposure. The synovial fluid alpha defensin test overdiagnosed infection in 4 patients with metallosis and 2 patients with only 2 minor MSIS criteria for periprosthetic joint infection. These 2 patients were treated with 2-stage reimplantations despite the MSIS data and had negative intraoperative cultures. Might organisms that were difficult to culture have been eradicated by the 2-stage revision? The MSIS criteria, although a brilliant contribution, are not the final word. Perhaps they were wrong. When should we order the synovial fluid alpha defensin? The current article includes an algorithm advising how to interpret the 4 permutations of synovial fluid alpha defensin and synovial fluid CRP. It does not explain when the test was indicated. This was a retrospective review of patients who had the test ordered by several physicians in varied circumstances. Future structured studies will be important. Should we order synovial fluid alpha defensin when the American Academy of Orthopaedic Surgeons guidelines indicate an aspiration (clinical suspicion or elevated erythrocyte sedimentation rate [ESR] and CRP)? If the synovial fluid is aspirated and if MSIS criteria define periprosthetic joint infection, then the cell count and differential are not yet likely to be replaced. Will the threshold values for synovial fluid alpha defensin and synovial fluid CRP prove to be more reliable than cell counts? Perhaps MSIS criteria need to be tested by long-term clinical studies. This would be challenging. If an MSIS-negative revision fails because of a periprosthetic joint infection, was that a new infection or a missed infection? In a 2017 lecture, “Prosthetic Joint Infection: A True Menace,” Dr. Parvizi made the point that even registry data must be regarded skeptically when it comes to infection5. Many cases reported as “aseptic” loosening are likely missed periprosthetic joint infection. Maybe arthroplasties are “all infected, all the time” and increased bacteria and the host ecology yield a periprosthetic joint infection? Maybe the immune system occasionally overresponds, producing alpha defensin prematurely as in metallosis? Perhaps certain bacteria multiply without provoking an immune response? If Mother Nature can be fooled, it is no surprise that our tests have limits too. It is a complicated game of numbers.