The Graphical Representation of Cell Count Representation: A New Procedure for the Diagnosis of Periprosthetic Joint Infections

Abstract

Aim: This study was designed to answer the question whether a graphical representation increase the diagnostic value of automated leucocyte counting of the synovial fluid in the diagnosis of periprosthetic joint infections (PJI).

Material and methods: Synovial aspirates from 322 patients (162 women, 160 men) with revisions of 192 total knee and 130 hip arthroplasties were analysed with microbiological cultivation, determination of cell counts and assay of the biomarker alpha-defensin (170 cases). In addition, microbiological and histological analysis of the periprosthetic tissue obtained during the revision surgery was carried out using the ICM classification and the histological classification of Morawietz and Krenn. The synovial aspirates were additionally analysed to produce dot plot representations (LMNE matrices) of the cells and particles in the aspirates using the hematology analyser ABX Pentra XL 80.

Results: 112 patients (34.8%) had an infection according to the ICM criteria. When analysing the graphical LMNE matrices from synovia cell counting, four types could be differentiated: the type "wear particles" (I) in 28.3%, the type "infection" (II) in 24.8%, the "combined" type (III) in 15.5% and "indeterminate" type (IV) in 31.4%. There was a significant correlation between the graphical LMNE-types and the histological types of Morawietz and Krenn (p < 0.001 and Cramer test V value of 0.529). The addition of the LMNE-Matrix assessment increased the diagnostic value of the cell count and the cut-off value of the WBC count could be set lower by adding the LMNE-Matrix to the diagnostic procedure.

Conclusion: The graphical representation of the cell count analysis of synovial aspirates is a new and helpful method for differentiating between real periprosthetic infections with an increased leukocyte count and false positive data resulting from wear particles. This new approach helps to increase the diagnostic value of cell count analysis in the diagnosis of PJI.

Keywords: cell count; diagnosis; leukocyte; periprosthetic joint infection.

Figure 1

LMNE matrix with the different fields for the leukocyte populations and the NOISE area.

Figure 2

LMNE matrix of a type I (abrasion type) with a cloud in the NOISE-area of a 65-year-old male patient with an aspirate of the hip arthroplasty 15 years postoperative. The measured cell count was 1500 cells/µL.

Figure 3

(a) LMNE matrix of a type I with polyethylene wear particles produced in a laboratory. The cloud is at the top in the NOISE area. (b) LMNE matrix of a type I with metal debris particles in a 73-year-old male patient with an articulation of a ceramic head on the inner side of a cup with disturbed inlay. The cloud is at the left bottom close to the NOISE area and the distribution is “L”-shaped. The measured “cell count” was 6700 cells/µL.

Figure 3

(a) LMNE matrix of a type I with polyethylene wear particles produced in a laboratory. The cloud is at the top in the NOISE area. (b) LMNE matrix of a type I with metal debris particles in a 73-year-old male patient with an articulation of a ceramic head on the inner side of a cup with disturbed inlay. The cloud is at the left bottom close to the NOISE area and the distribution is “L”-shaped. The measured “cell count” was 6700 cells/µL.

Figure 4

LMNE matrix of a type II (infection type) with a cloud in the area of the neutrophil leukocytes in a 75-year-old patient with a late periprosthetic joint infection of a total knee arthroplasty. The measured cell count was 1840 cells/µL.

Figure 5

LMNE matrix of a type III (combined type) with one cloud in the area of the neutrophil leukocytes and a second cloud in the NOISE area in a 76-year-old male patient with a periprosthetic joint infection of a total knee arthroplasty. The measured cell count was 5840 cells/µL.

Figure 6

LMNE matrix of a type IV (indifference type) with no clear cloud or increase in cell types or particles in a 73-year-old patient. The measured cell count was 240 cells/µL.

Figure 7

Receiver operating characteristics curve (ROC-curve) with the calculation of the threshold of cell count at a value of 1400 cells/µL with a sensitivity of 90.2% and specificity of 91.9%.