Late and/or chronic periprosthetic joint infection in patients with total joint replacements can be very difficult to diagnose, and evaluation of the levels of serum inflammatory markers is one of the more commonly done diagnostic tests when periprosthetic joint infection is considered. Unfortunately, inflammatory arthritis, in its many forms, is a common etiology of joint destruction leading to arthroplasty, results in mild or moderate degrees of immunocompromise, and is associated with persistent, baseline elevations in levels of inflammatory biomarkers. Because of this terrible triad, patients with inflammatory arthritis may be at greater risk of chronic periprosthetic joint infection, and such infections may be harder to diagnose because elevations in inflammatory markers may reflect the underlying disease rather than infection of the prosthetic joint.
Until now, very little actual information has existed to aid clinicians who face the difficult task of trying to diagnose an infection at the site of a joint replacement in a patient with known inflammatory arthritis. In fact, a workgroup consisting of members of the Musculoskeletal Infection Society recently proposed a new definition of periprosthetic joint infection, but none of the studies on which they based their work included patients with underlying inflammatory diseases1. Thus, a number of important clinical questions remain unanswered. During a workup of an infection in such a patient, are mild or moderate elevations in C-reactive protein (CRP) diagnostic of a deep infection or are they just a reflection of the systemic illness? Similarly, can synovial leukocyte levels be relied on? Do these otherwise standard diagnostic markers have any role in this setting? Despite possible baseline elevations in inflammatory proteins, are there reliable thresholds for inflammatory markers or leukocyte counts above which periprosthetic joint infection can be considered to be present in patients with inflammatory disease?
In this issue of The Journal of Bone and Joint Surgery, Cipriano et al. provide data on levels of serum erythrocyte sedimentation rate (ESR), serum CRP, and synovial fluid white blood-cell count with differential in a cohort of 803 patients undergoing 871 consecutive hip and knee arthroplasties (including sixty-one in patients with inflammatory arthritis). Patients with true infection were identified on the basis of a combination of culture results and histopathological findings. Receiver operating characteristic curves were prepared for all three parameters and used to establish optimal thresholds for the diagnosis of periprosthetic joint infection.
Not surprisingly, there was a significantly higher rate of infection following procedures in patients with inflammatory arthritis than that following procedures in patients with noninflammatory arthritis (31% compared with 18%, p = 0.013). However, on the basis of their receiver operating characteristic curve analysis, the authors concluded that these serum (ESR and CRP) and synovial tests performed similarly in both groups of patients, with similar diagnostic thresholds, area under the receiver operating characteristic curve, and calculations of sensitivities, specificities, negative predictive values, and positive predictive values. Furthermore, the authors concluded that the synovial leukocyte count with differential performed better than the serum inflammatory markers for all patients.
This study by Cipriano et al. is important because it provides some of the first data regarding the ability of these common diagnostic tests to discriminate periprosthetic joint infection in patients with an underlying inflammatory disease. It is a simple conclusion to remember that the tests performed similarly in both groups of patients and that the optimal diagnostic thresholds are the same that surgeons are accustomed to using.
However, there are some limitations to this study of which the reader needs to be aware. The time when the preoperative serum markers were obtained was not standardized with respect to the date of surgery; they were obtained at the time of the first consultation with the surgeon, and the interval between this visit and the date of surgery was variable. The same concern applies to the joint aspirations; some were done at the time of the preoperative visit and others were done the day of surgery. Although the timing of these tests reflects clinical practice and makes the study “real,” it could be a problem if these values change during the varying interval between when they were obtained and when surgery was performed. It would be of interest to determine whether laboratory tests performed at a standardized time, relatively close to surgery, would have different results and, therefore, different diagnostic thresholds. When one compares the calculated sensitivities and specificities of the tests, one sees that, in general, the specificity was lower and the sensitivity was higher in the patients with inflammatory disease. However, as noted by the authors, the prevalence of infection was different in the two groups of patients, affecting the positive and negative predictive values.
For now, it appears that the same approach that is used for the diagnosis of chronic periprosthetic joint infection in patients with noninflammatory arthritis can be used in patients with inflammatory disease. The data presented by Cipriano et al. suggest that the specificity of these tests is lower in patients with inflammatory arthritis than in those with noninflammatory arthritis, meaning that these tests are less likely to be normal in patients without periprosthetic joint infection. Because of the higher prevalence of infection in such patients, the predictive values of the tests are similar. The diagnosis of periprosthetic joint infection remains difficult, and clinicians should be aware of the potential pitfalls of diagnosis in patients with underlying systemic disease.