When Is a Positive Culture in Shoulder Surgery Not an Infection? Commentary on an article by William R. Mook, MD, et al.: "The Incidence of Propionibacterium acnes in Open Shoulder Surgery. A Controlled Diagnostic Study".

Abstract

Commentary Propionibacterium acnes represents the most common pathogen in periprosthetic joint infections of the shoulder, accounting for up to 70% of positive culture results at the time of revision surgery1. This diagnosis can be elusive for several reasons. P. acnes is a commensal organism of low virulence that populates the dermal sebaceous glands about the shoulder2. Its clinical presentation is often insidious and delayed and is frequently characterized clinically by persistent shoulder pain and a paucity of cutaneous, systemic, or radiographic signs. Furthermore, the preoperative work-up has been characterized historically by poor sensitivity. Acute-phase serologic markers such as the erythrocyte sedimentation rate and C-reactive protein level are often normal in P. acnes infections and the organism can be difficult to culture, requiring long-duration culture holds in multiple culture media3 for optimal culture yield. The usefulness of intraoperative frozen sections for aiding in the diagnosis of periprosthetic joint infections of the shoulder varies from one pathology laboratory to another, but has been inconsistent at best4. Because of the low virulence of P. acnes and the associated diagnostic delay and uncertainty, surgeons are often confronted with difficult treatment dilemmas at the time of revision surgery, such as determining whether or not a revision should be carried out in a staged fashion until a periprosthetic joint infection can be effectively ruled out. In addition, positive cultures at the time of revision surgery may represent a true periprosthetic joint infection but may also arise unexpectedly as a contaminant during specimen procurement and culture incubation. Mook et al. set out to determine the rate of positive culture growth from deep-tissue specimens obtained from patients undergoing open surgery of the glenohumeral joint as well as the demographic correlates of the culture results. One of the study strengths is the availability of a control group comprising cultures of sterile gauze sponge specimens. The authors found that male sex strongly correlated with positive P. acnes cultures, which is in keeping with previous studies1,5. The authors also found that two or more previous intra-articular corticosteroid injections correlated with positive P. acnes cultures, which is compelling, given the frequency with which these are performed prior to the consideration of prosthetic shoulder arthroplasty. However, the authors also acknowledged that post hoc analysis revealed that their study was underpowered to test this association. The most striking study finding is the 13% prevalence of positive culture growth from the control gauze specimens. In addition, 18% of patients without prior surgery had at least one positive culture, 93% of which were positive for P. acnes, despite standardized and meticulous sample procurement. Finally, the authors observed that cultures continued to grow out over time, suggesting a greater prevalence of contamination after fourteen days of incubation. Taken together, these findings suggest the principal culprit for these false-positive cultures is sample contamination along the chain of custody from sample procurement to culture incubation. However, positive intraoperative cultures may also arise following inadvertent contamination during sample procurement from the dermal layers of the shoulder incision. Recently, Matsen et al. found that a nearly identical 14% of deep culture specimens were positive for P. acnes despite timely administration of intravenous antibiotics in a series of ten male patients undergoing primary shoulder arthroplasty6. In addition, Levy et al. observed a 42% prevalence of positive cultures at the time of primary arthroplasty in patients with no prior surgery7. This discrepancy in prevalence may be explained in part by the fact that, in the study by Levy et al., intravenous antibiotics were withheld until after culture specimens had been obtained7, whereas in the current study, these were administered preoperatively. Collectively, these data demonstrate that not every positive culture represents a true infection and raise concerns about the use of long-hold cultures as a gold standard. At the very least, the duration of incubation must be optimized to balance maximizing culture sensitivity and specificity. The concerns raised in this study regarding positive P. acnes cultures provide an impetus for the development of alternative diagnostic strategies that may improve preoperative and intraoperative evaluation of periprosthetic joint infection in the shoulder. The pro-inflammatory cytokine interleukin-6 (IL-6) in both serum and synovial fluid has shown increased sensitivity and specificity in the diagnosis of periprosthetic joint infection of the hip and knee8,9. Its use as a serum marker in periprosthetic joint infection of the shoulder has not shown similar findings, with very low sensitivity in cases positive for P. acnes and other low-virulence organisms10,11. In contrast, elevated levels of IL-6 in synovial fluid have shown promising results. A recent study evaluating the diagnostic utility of synovial fluid IL-6 levels in revision shoulder arthroplasty reported a sensitivity of 87% and a specificity of 90% for the diagnosis of periprosthetic joint infection, with nearly all cases culture-positive for P. acnes or another indolent organism12. Synovial fluid IL-6 levels were significantly elevated in P. acnes-positive culture cases and in one-stage revisions with no preoperative or intraoperative findings of infection but with multiple positive intraoperative culture results. Other synovial fluid biomarkers have been investigated with promising results, including alpha-defensin, an antimicrobial peptide primarily expressed in polymorphonuclear cells13. In another recent study on diagnostic utility in periprosthetic joint infection of the shoulder with indolent organisms, synovial fluid alpha-defensin demonstrated a sensitivity of 63% and a specificity of 95% with significantly increased levels in revision cases with P. acnes-positive cultures14. The further development of these and other synovial fluid biomarkers to facilitate preoperative and intraoperative use may lead to better decision-making in the management of periprosthetic joint infection of the shoulder, including the decision to proceed with one-stage revision compared with two-stage revision or the need for postoperative antibiotic therapy in the setting of an unexpected positive culture result. In addition, recent data suggest that the use of alternate criteria for intraoperative frozen section analysis may improve the ability to detect P. acnes infections, with sensitivity increased to 72% and specificity maintained at 100% using a threshold of more than ten polymorphonuclear cells in each of the five most dense high-power fields15. Although a precise algorithm does not currently exist for the diagnosis of periprosthetic joint infection of the shoulder, including P. acnes-positive infections, interpretation of a positive culture result should be made in the context of the overall clinical picture. This should take into account other positive preoperative and intraoperative markers for infection, including traditional serum markers and intraoperative frozen section findings, as well as newer synovial fluid biomarkers and the characteristics of the positive culture result(s) themselves, such as the timing of the first positive culture and the number of positive culture results relative to the overall number of cultures taken6. Such data taken together can help to determine whether a positive culture is likely to represent a false-positive result consistent with contamination or a true-positive finding. This study and others highlight the dearth of precise algorithms for both diagnosis and treatment of periprosthetic joint infection of the shoulder. Essential to such algorithms is the development of a consensus definition of periprosthetic joint infection of the shoulder, based on a combination of preoperative and intraoperative findings and intraoperative culture results. Recent studies have attempted to interpret the results of newer diagnostic tests within the framework of such criteria11,12,14,15. However, the evaluation and management of the painful shoulder arthroplasty remain highly variable and need to be standardized in areas such as preoperative surgical site preparation, choice and timing of intraoperative antibiotics during revision surgery, number and type of intraoperative cultures obtained during revision surgery7,16, culture methods and length of time for culture holds, and choice and length of postoperative antibiotic therapy. A consensus definition of periprosthetic joint infection of the shoulder and a standardized approach to the evaluation and management of the painful shoulder after arthroplasty will aid in interpreting research data and developing more robust diagnostic algorithms and successful clinical treatment pathways in the future.