[Validity of the Morawietz classification for evaluation of periprosthetic tissue]

PURPOSE OF THE STUDY A consensual classification of the periprosthetic interface membrane obtained at revision total joint arthroplasty was published by Morawietz et al. in 2006. Based on histomorphological criteria, four types of periprosthetic membrane were proposed: type I, aseptic failure; type II, septic failure; type III, combined type (carrying signs of both type I and II); and type IV, indeterminate type. The aim of this study was to find out whether and to what extent the Morawietz system would be suitable for use at an independent institution involved in the evaluation of periprosthetic membranes for a long time. Should it appear that the institution achieved an equally good or even better agreement between the clinical diagnosis and the histopathological finding, this consensus classification could be recommended for routine use. MATERIAL AND METHODS The samples of periprosthetic tissue evaluated in this study were obtained during surgery from the following groups of patients: 66 patients with aseptic loosening of total hip (THA) or knee arthroplasty, 15 patients with infection of THA, 16 patients with THA without any signs of aseptic loosening, osteolysis or infection; 8 patients with hip osteoarthritis and 8 patients with knee osteoarthritis. Sample collection and processing (for purposes of histomorphological evaluation and immunohistochemical staining) was performed according to the established protocol. The tissue samples evaluation was made by an experienced pathologist hand in hand with the method described in the original paper by Morawietz et al. For a more detailed tissue analysis, selected antibodies (CD4, CD8, CD20, IFN-γ and Hsp-60) were visualized by immunohistochemistry. RESULTS The majority of samples from aseptic reoperations were classified as membranes of the type I (79%) and III (16%). Specimens retrieved from septic cases were mostly classified as membranes of type II and III (60% together). The septic membranes showed a significantly higher expression of CD20 protein when compared with both the aseptic (p < 0.0001) and control THA samples (p = 0.003). The membranes retrieved from the surroundings of a stable THA without osteolysis and infection had lower expression levels of Hsp60 and IFN-γ, when compared with those from both aseptic and septic loosening. Finally, Hsp-60 expression was significantly higher in osteoarthritic tissue than in samples from stable THA (p = 0.041). DISCUSSION Morawietz et al. proposed a standardized classification system for evaluation of periprosthetic tissue. As any attempt at generalization of a complex issue, this proposal has certain shortcomings. One of these is poor detection of chronic and low-grade infections. A method that would improve the conventional counting of polymorphonuclear leukocytes is still being sought. In this connection, immunostaining for CD20 combined with an assessment of antimicrobial peptides may be a promising option. The supplementary specimen staining showed that pseudosynovial tissue is much more active in patients carrying infection and the least active in samples from stable THA in which certain tolerance and thus tissue homeostasis might be expected. CONCLUSIONS 1. In this study the distribution of findings classified according to the Morawietz system was similar to the results published in the original study from 2006. 2. The definition of an aseptic membrane (type I) in the Morawietz system meets the requirements of clinical practice (agreement, about 80%). 3. An increased sensitivity for infectious membrane detection can be achieved by using supplementary immunohistochemical staining effective particularly in chronic and low-grade infections. 4. Painless and stable THAs typically have very low expression levels of CD4, CD20 and Hsp-60 proteins, and interferon- -gamma (IFN-γ) as well. Key words: total hip arthroplasty, total knee arthroplasty, aseptic loosening, prosthetic joint infection, tissue analysis, membranes, CD receptors, Hsp-60 protein, IFN-γ.