Abstract: OBJECTIVE: There is still no single indicator or even a combination of indicators that can definitely confirm or exclude periprosthetic joint infection. D-dimer has recently emerged as a novel index for the diagnosis of periprosthetic joint infection, but its reliability is uncertain. We determined the practical application value and the accuracy of D-dimer in the diagnosis of periprosthetic joint infection by using meta-analysis.
METHODS: We conducted a systematic search and screening to identify relevant articles from register databases such as PubMed, Embase, Web of Science, and the Cochrane Library. QUADAS-2 was used to assess the quality of each included study. The bivariate mixed-effects regression model was used to combine sensitivity, specificity, likelihood ratio, diagnostic odds ratio, summary receiver operating characteristic curve, and area under summary receiver operating characteristic curve to evaluate the diagnostic value of overall D-dimer for periprosthetic joint infection. Univariate meta-regression and subgroup analysis were performed to explore the sources of heterogeneity.
RESULTS: (1) A total of 11 studies were included in our study, including 1 645 patients. The quality of the included studies was regarded as the upper-middle level. (2) The pooled sensitivity, specificity, pooled positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio, and area under the curve were 0.80(95%CI:0.72-0.87), 0.74(95%CI:0.62-0.83), 3.04(95%CI:2.01-4.61), 0.27(95%CI:0.18-0.40), 11.40(95%CI:5.50-23.63), and 0.84(95%CI:0.81-0.87). (3) Univariate meta regression showed that sample source and threshold had influence on the heterogeneity of pooled sensitivity results. Subgroup outcomes showed that serum D-dimer had a higher diagnostic accuracy than plasma D-dimer for periprosthetic joint infection (pooled sensitivity: 0.88 vs. 0.67; pooled specificity: 0.78 vs. 0.67; pooled area under the curve: 0.91 vs. 0.68). (4) The Spearman correlation analysis showed that there was no obvious threshold effect among the studies. The heterogeneity might be unrelated to the threshold effects.
CONCLUSION: (1) Serum D-dimer had a promising performance for the diagnosis of periprosthetic joint infection, which was shown to have a better diagnostic value than plasma D-dimer. (2) There are no sufficient data existing for subgroup analysis of different periods or joints of periprosthetic joint infection, so we could not assess the differences of D-dimer in diagnosing acute periprosthetic joint infection versus chronic periprosthetic joint infection, or hip joint infection versus knee joint infection. Thus, more large, multicenter prospective studies are needed for verification. 

Key words: bone, hip joint, knee joint, periprosthetic joint infection, D-dimer, arthroplasty, diagnosis, meta-analysis