Is D-dimer a Reliable Serum Marker for Shoulder Periprosthetic Joint Infection?

Abstract

The diagnosis of periprosthetic shoulder infection continues to be difficult to make with confidence. Serum D-dimer has proven to be effective as a screening tool for periprosthetic joint infection in other major joints; however, it has yet to be evaluated for use in periprosthetic shoulder infection. (1) Is D-dimer elevated in patients with probable or definite periprosthetic shoulder infections? (2) What is the diagnostic accuracy of D-dimer for periprosthetic shoulder infections? (3) What are the diagnostic accuracies of serum tests (erythrocyte sedimentation rate [ESR], C-reactive protein [CRP], and D-dimer), singly and in combination? Between March 2016 and March 2020, 94 patients undergoing revision total shoulder arthroplasty (anatomic or reverse) at a single institution had preoperative serum testing with CRP, ESR, and D-dimer. These 94 patients were a subset of 189 revision shoulder arthroplasties performed at this institution during the study period who met inclusion criteria and consented to participate. Included patients had a mean ± SD age of 69 ± 8 years, and 56% (53 of 94) were men. Patient records were reviewed to classify patients as definitely having infection, probably having infection, possibly having infection, or unlikely to have an infection, according to the International Consensus Meeting (ICM) definition of periprosthetic shoulder infection. Statistical analyses, including a receiver operating characteristic curve analysis, were performed to quantify the diagnostic value of D-dimer for periprosthetic shoulder infection. Based on the ICM definition, 4% (4 of 94), 15% (14 of 94), 14% (13 of 94), and 67% (63 of 94) of patients had definite, probable, possible, or unlikely periprosthetic shoulder infections. D-dimer was elevated in patients with definite or probable infections (median [range] 661 ng/mL [150 to 8205]) compared with those with possible infections or those who were unlikely to have an infection (263 ng/mL [150 to 3060]; median difference 143 ng/mL [95% CI 40 to 503]; p = 0.01). In the receiver operating characteristic curve analysis, D-dimer had an area under the curve of 0.71 (0.50-0.92), demonstrating weak diagnostic value. A D-dimer level of 598 ng/mL provided a sensitivity and specificity of 61% (95% CI 36% to 82%) and 74% (95% CI 62% to 83%), respectively, for diagnosing a definite or probable infection according to the ICM definitions. The specificity of detecting periprosthetic joint infection (88% [95% CI 79% to 94%]) was high when three positive serum markers (ESR, CRP, and D-dimer) were required, at the expense of sensitivity (28% [95% CI 10% to 53%]). In periprosthetic shoulder infection, D-dimer is elevated. However, similar to other serum tests, it has limited diagnostic utility in identifying patients with periprosthetic shoulder infection. Further work is needed to understand the process by which D-dimer is associated with active infection. Level III, diagnostic study.