The aim of this study was to investigate the accuracy of pre- and intraoperative diagnostic parameters of periprosthetic hip joint infection. Additionally a strategy of preoperative selection of patients with high suspicion of PJI was evaluated and compared with the final postoperative diagnosis.
A limitation of this study is that the study has no control group. This limitation is caused by the fact that there is a lack of a gold-standard definition of periprosthetic joint infection in the current literature. Periprosthetic joint infection is a multimodal process based on different causes and occurs in a variety of clinical presentations. Different diagnostic parameters are published with a broad range of sensitivity and specifity. The Diagnosis of PJI depends on several tests rather than on a single test. An additional problem is that some tests are only postoperatively available such a microbiological cultures or histopathological evaluation. Therefore the preselection of patients with a high suspicion of periprosthetic joint infection is a vast challenge.
In this study the histopathological investigation turns out as a very practical and valid diagnostic tool for intraoperative detection of periprosthetic joint infection with a high sensitivity (0.95) and specificity (0.92). Because of detailed histopathological and polarised characterisation of the periprosthetic interface membrane the clarification whether loosening is due to bacterial infection or not is very precise. A harvesting of tissue samples was possible in all cases. Caused by the study design, tissue samples were only taken in patients with clinical or anamnestic suspicion of infection. A recently published study by Morawietz et al, in which 370 periprosthetic membranes from revision surgery were analyzed, could be shown, that most of the samples (94.9%) were suitable for histological classification [3
]. A differentiation of infected and non-infected loosening was well possible. A discrepancy between microbiological and histological findings was found in only 10.7% of the cases.
In 28 of the 37 septic prostheses (75%) the histological and microbiological results were concordant. Similar investigations found a concordance of 89 percent (155/174) [3
]. This fact raises the question as to whether microbiological culture or histological examinations are more valid with respect to sensitivity and specificity. Both tests have there individual pitfalls as inappropriate incubation time, previous antimicrobial therapy given to the patient, contamination in terms of cultures or e.g. insufficient preparation of the tissue samples.
To improve the results of histological diagnosis, the surgical pathologist should be provided with additional clinical data, such as the lifetime of the prosthesis, type of fixation, relevant records on clinical pathology, and microbiological findings by the orthopaedic surgeon. This information would help the pathologist interpret results of histopathological samples. Caused by the necessity of tissue sample preparation, an intraoperative statement of the pathologist was not possible in this study. It should be proofed whether the classification system of periprosthetic interface membrane is applicable to frozen section or preoperative biopsy of the neocapsule because a pre- or direct intraoperative test result would be a worthwhile effort. Despite the fact, that the neocapsule is not responsible for loosening, it is generally accepted that the changes in histological appearance are very similar in these different tissue specimens in the same patient caused by interaction of the new joint space with the periprosthetic space [12
Intraoperative cultures are a crucial parameter in diagnosis of PJI and therefore cultures are frequently used as the gold standard to which every other diagnostic parameter is correlated [2
]. Without correct detection and identification of microorganisms the final diagnosis is ambiguous and adequate antibiotic treatment may not be realized. However, in literature intraoperative cultures have a broad range of sensitivity (range 0.65 to 0.94 (0.78 in this study)) and specificity (range 0.71 to 1.0) (0.92 in this study)) depending on the definition of infection and they are subjected to a variable rate of false positive and negative results [4
]. In this study tissue cultures yielded a false result in 8 (16%) of 50 patients (seven false-negative results (14%) and 1 false-positive result (2%)) what is similar to results reported in the literature [12
]. Inadequate incubation time, inappropriate choice of media and antimicrobial therapy, as well as sample contamination from human skin flora are responsible for false-negative or false-positive results. Such problems reduce the level of significance of microbiological culture methods, and have been pointed out in several studies [12
]. It has been reported that, due to the small numbers and low metabolism of bacteria involved in periprosthetic infections generally increase the time needed to resuscitate them [17
]. Therefore, the growth period should be extended to increase the detection rate of infectious bacteria in excised tissue samples. It could be shown that some microorganisms require a minimum incubation time of 8 days, since these microorganisms grow slowly [10
Furthermore, it has been emphasized, to improve the hospital culturing of tissue samples, at least eight tissue samples should be taken from different sites in the operative field [19
]. Recent studies criticise that traditionally standard diagnostic tests are designed for examining planktonic bacteria and those that are adequate for detecting of sepsis-related pathogens without involvement of foreign material; therefore a significant number of infections of orthopaedic devices may remain undetected [9
]. The nutrient media and isolation procedures do not provide the requisite conditions for recovering such bacteria in culture.
In this study, preoperative hip aspiration had a low sensitivity (0.57) and specificity (0.5), indicating that indolent joint infection cannot be diagnosed on the basis of aspiration alone. However, it may be the most suitable preoperative tool to provide preoperative information, such as the identity of the infecting organism and it sensitivity to antibiotics [20
]. In the literature, preoperative joint aspiration for detecting PJI has a broad range of values of sensitivity varying between 0.11 and 1.00 and specificity varying from 0.78 to 1.00 [2
] certainly depending on the different technique or definition of infection.
Most critical issue in hip aspiration is a high false-positive 6/50 12% and false-negative 14/50 28% rate due to contamination at the time of aspiration or in the microbiology laboratory or false-negative rate due to low concentrations of organisms, delay in transport or inoculating the sample. The inability to aspirate fluid and subsequent washout with saline may contribute to samples with low concentration.
Microorganisms involved in infections of orthopaedic devices are highly adapted on the implant or in the bone-cement interphase, adhering to the environment of the in vivo biofilm, but not planktonic in the synovia and therefore join aspiration may be insufficient [9
Our results demonstrate that hip aspiration is only of assistance if there is any conspicuous preoperative sign such as an open wound or sinus in communication with the joint or in case of elevated C-reactive protein which could not be explained by other conditions what is corresponding to the experience of other authors [2
An elevated C-reactive protein which could not be explained by other conditions highly indicates a PJI especially if there is evidence from clinical or radiographic examination.
In the literature, values of sensitivity and specificity range from 0.61 to 1.0 and from 0.81 to 1.0 [2
]. In general, C-reactive protein is a relevant parameter in diagnosing PJI and the elevation of C-reactive protein is a prerequisite to joint aspiration.
Another recommended blood parameter in detection of periprosthetic hip joint infection is the erythrocyte sedimentation rate. In this study we focused our investigation only on the C-reactive protein. Our decision to concentrate on C-reactive protein was reinforced by the fact that the C-reactive protein level increases from normal values to reach maximum values within 24 hours after surgery and then returns to trace amounts in approximately two to three weeks [25
]. The erythrocyte sedimentation rate may remain elevated for months after an uncomplicated total hip replacement [24
]. Therefore, the ability of the C-reactive protein level to return to normal much faster than the erythrocyte sedimentation rate enables it to be a more sensitive indicator of infection, particularly in the early postoperative period.
Definitely postoperative infection could not be confirmed in 13 patients, although there were clear preoperative findings which highly indicated PJI. None of them had an open wound, a fistula or purulent aspiration.
In 6 of these 13 patients the preoperative suspicion of PJI was reinforced by a positive preoperative hip aspiration. But intraoperative microbiological cultures and histopathological results could not confirm the positive results of preoperative aspiration at these patients. It has to assume that aspiration results were false positive.
Also a two stage revision was performed in the remaining 7 patients even though preoperative aspiration was negative. Conspicuous clinical signs like previous septical revision (3 cases), early failure within 5 years (n = 3), and unclear elevated C-reactive protein in connection with persistent hip pain (n = 1) led to septical revision procedure as a precaution not to overlook a creeping infection.
Obviously, there is a difference between the preoperative suspicion of PJI (depending on the clinical presentation and the evaluation of the consultant) and the final postoperative diagnosis that could be made by consideration of all pre- and intraoperative test results. This raises the question of whether the consultant is too prudent in diagnosing PJI or the final definition of PJI is not precise enough or the diagnosis parameters are too insensitive. In consideration of the fact that a miss diagnosed periprosthetic joint infection may have serious consequences for the patient, a minimum of over diagnosed and over treatment is tolerable. Moreover, in case of negative intraoperative test results and absence of fistula, open wound or purulent aspiration, reimplantation would be performed earlier (as soon as test results come up) than it is usual in the standard two stage revision procedure. Negative effects of a two stage revision like muscles atrophy, immobilisation, contractures and leg length differences will be limited to minimum. On the other hand, the relative high number of patients drop out of the diagnosis pattern, could be explained through the definition of PJI. For definition of PJI we tried to include both, clinical presentation and pre/intraoperative test results, similar to the definition of Spangehl et al and Giulieri et al, to obtain a diversified diagnostic pattern [4
]. But nevertheless, all tests are subjected to a certain rate of false positive and false negative test results, as this and other studies have shown. To obtain highest accuracy in diagnosis of PJI and to improve the detection rate, an exact implementation and interpretation of each single diagnostic test is crucial. Improvements in the technique of joint aspiration, reported by Ali et al, appropriate incubation time of the cultures, the correct choice of media, and stopping previous antimicrobial therapy in advance, are all important points [6
]. However, it is recommended that joint aspiration should not be performed, if there is no elevation of C – reactive protein, or clear clinical or radiological signs. Other reasons for elevation of C – reactive protein should be excluded.
Recent investigations have pioneered new techniques for detection of PJI. A notable success in detection of prosthetic hip infection at revision arthroplasty could be achieved by Immunofluorescence Microscopy, Polymerase Chain Reaction (PCR), by analysis of explanted prostheses surfaces and by confocal laser scanning microscopy [18
]. However, it remains to be seen if these new techniques can be established in clinical practice.