Although etanercept, a dimeric fusion protein linking two 75-kDa tumor necrosis factor (TNF)–a receptor molecules to the Fc portion of immunoglobulin G1, is clearly effective for inhibiting TNF-mediated synovial inflammation and joint destruction in patients with rheumatoid arthritis [1, 2], the role of anti-TNF adjuvant therapy for septic arthritis is far from certain. Because proinflammatory mediators such as TNF-a play an important role in the pathogenesis of septic arthritis [3], there is reason to believe that antiTNF therapy in combination with antibiotics may be beneficial. Two other studies have demonstrated that systemic corticosteroids or a nonsteroidal antiinflammatory drug administered in combination with antibiotics was superior to antibiotics alone in animal studies of Staphylococcus aureus septic arthritis [4, 5]. A double-blind, placebocontrolled randomized clinical trial also demonstrated that short-course dexamethasone adjuvant therapy significantly reduced symptom duration and joint dysfunction in children with hematogenous septic arthritis, including those with S. aureus infection [6]. In this issue of the Journal, Fei et al [7] provide evidence suggesting that anti-TNF adjuvant therapy may also be beneficial in S. aureus septic arthritis and sepsis. Using a murine hematogenous infection model, they demonstrated significant reduction in both clinical severity (arthritic index) and bone destruction (histopathologic index) after combination treatment with cloxacillin and etanercept, compared with cloxacillin alone. Furthermore, overall mortality at 14 days was reduced in mice treated with the combination. However, this study also raises a number of questions. Is this animal model relevant to human disease? Is the favorable response of etanercept specific to the strain of S. aureus used in their animal model? Is there sufficient evidence to warrant further studies of etanercept by randomized clinical trials in patients with S. aureus septic arthritis? What about safety concerns with etanercept? An important but unexplained observation in their murine model of S. aureus septic arthritis is the high early mortality rate in mice receiving antibiotics alone ( 65% within 2 days after treatment). In contrast, mice receiving phosphate-buffered saline (PBS) had a more gradual death rate ( 45% at 10 days after infection). Fei et al suggested that this early mortality in the antibiotic treatment group could be due to the release of cell wall components after exposure to bactericidal antibiotics. This is certainly a possibility, because peptidoglycan and lipoteichoic acid are released from grampositive cell wall during antibiotic exposure, and are known to activate a cascade of proinflammatory cytokines, including TNF-a, interleukin-1, and interleukin-6 [8, 9]. However, whether release of these bacterial cell wall components contributes to early mortality after antibiotic treatment during experimental S. aureus infection is unknown. An alternative and equally attractive proposition is that this early mortality was due to the release of superantigen exotoxins from the infecting strain of S. aureus after treatment with bactericidal antibiotics [10], because LS-1, the S. aureus strain used in the experimental model of septic arthritis, is known to produce toxic shock syndrome toxin 1 (TSST-1) [11, 12]. TSST1 was found to be both arthritogenic and lethal in this murine model, because TSST-1–deficient mutants were significantly less arthritogenic and produced less severe disease than their isogenic parental strains [13]. Thus, the beneficial effect of etanercept in this animal model was probably derived from its anti-TNF effect on TSST-1–mediated lethality, as confirmed by its protective Received 9 March 2011; accepted 14 March 2011. Potential conflicts of interest. none reported. Correspondence: Anthony W. Chow, MD, 769 Burley Pl, West Vancouver, BC, CanadaV7T 2A2 (tonychow@interchange.ubc.ca). The Journal of Infectious Diseases 2011;204:332–34 The Author 2011. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals. permissions@oup.com 0022-1899 (print)/1537-6613 (online)/2011/2043-0002$14.00 DOI: 10.1093/infdis/jir272