Diagnosis of total knee arthroplasty (TKA) infection relies on the combination of clinical findings, imaging and laboratory studies (from blood and synovial fluid). This latter group includes serum biomarkers, molecular testing, histology and especially microbiological techniques. Not all diagnostic tests are equally important for the diagnosis. In such a complex field where false positives and false negatives are frequent, a hierarchy is needed to understand the available evidence of infection in a specific case that would then inform the specific treatment.

A major accomplishment of the community in charge of patients suffering from TKA-related infection (in the orthopaedic field as well as in the infectious disease field) has been the international agreement on the definition of prosthetic joint infection (PJI), as reviewed by Tande and Patel [1].

Major criteria to provide definitive evidence of periprosthetic joint infection (PJI) include sinus tract communicating with the prosthesis and identical microorganisms isolated from two or more cultures. Purulence about the prosthesis may be definitive or supportive. Consensus also exists about the supportive (although not definitive) evidence provided by acute inflammation upon histological examination of periprosthetic tissue, single culture with any microorganism, single culture with a virulent microorganism, elevated synovial fluid leukocyte count, elevated synovial fluid neutrophil percentage and elevated serum erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) values. However, these diagnostic criteria on a painful TKA only permit to diagnose the case as potentially infected. Some of these cases are considered infected even if no microorganism is identified. Without the etiologic diagnosis, the treatment may not be specific enough to eradicate the infection. To establish the etiological diagnosis of a TKA infection, identifying the causative microorganism together with its antibiotic susceptibility is a key diagnostic issue to provide the correct treatment [2].

Samples used to investigate this causative microorganism are usually from the synovial fluid and/or the periprosthetic tissues. In a recent European survey about surgical diagnosis of infected TKA [3], 93.6% surgeons treating infected TKA and answering the survey routinely obtained tissue samples at surgery. Those were sent for culture (98.8%) and histology (72.5%). In this survey, synovial fluid exams included microbiology in 97.7% of cases.

Patients with culture-negative PJI have a diagnosis of infection without microbiological evidence. Periprosthetic purulence, acute inflammation determined by histopathology or a sinus tract communicating with the joint, without identifying causative microorganisms, can also be indicative of an infection, and the problem posed to the clinician in charge of the patient is significant [4]. This lack of microbiological confirmation may be due to prior antimicrobial therapy [5], or due to inadequate microbiological sample collection and processing at surgery, or also due to the inability of current methods to detect the pathogen (rare microorganisms or low-grade infections). The frequency of culture-negative PJI varies from 5% to 12% [4], and large centres with 7% culture-negative PJI tend to consider it a low incidence [6].

To increase the chances of culturing the causative microorganisms of prosthetic infection, its presence on the infected implant could offer the appropriate sample source to identify it.

During the joint colonization process that gives rise to a PJI, inert biomaterials are more prone than living tissues to suffer from bacteria adherence, according to the “race for the surface” theory of Gristina [7]. Adhered bacteria on the biomaterials, without suffering from the immunological response found in living tissues, may develop an extracellular matrix with specific characteristics called biofilm that protects the microorganisms from the activity of antimicrobials and from the surrounding host defence mechanisms [8]. Biofilm not only affects treatment, frequently requiring the explantation of the prosthesis to eradicate infection, but also limits microbiological diagnosis. Bacteria firmly attached to the implant and protected by the biofilm cannot be easily sampled and cultured. Swabs may just collect sessile bacteria in the surrounding fluids, while scraping with blades frequently fails to dislodge biofilm from metals [9]. Although it became evident that causative microorganisms attached to the prosthetic components offered the best opportunity to be detected when the components are explanted at revision surgery, the processing of the retrieved implant in the microbiology laboratory was cumbersome. To culture the whole implant was not practicable due to its size and potential contamination, and other techniques to dislodge microorganisms from implants had to be considered.

The use of ultrasound to release bacteria attached to prostheses was included in the ISO (International Standardization Organization) standards. This technique consists of low-frequency ultrasound waves passing through the liquid around the prosthesis and creating areas of high and low pressure. Bacteria are liberated from the surface of the implant when microscopic bubbles that are formed during the low-pressure stage collapse during the high-pressure stage. The fluid surrounding the implant with the dislodged bacteria can be submitted for culture or analysis.

Different sonication protocols have been developed in an effort to obtain the highest possible numbers of viable bacteria from the implant surface. In this context, the effects of the temperature, duration, composition of the sonication buffer and material in the sonication tube during bacteria sonication were evaluated [10], using sonication for 7 min at room temperature (22 °C). Further concentration of the sonicate solution by centrifugation allows a final volume of 400 μl to be obtained, and the suspended pellet can be divided into four portions of 100 μl for culture in different agar media. The duration of the sonication could be important to isolate Gram-negative bacilli, which can be eradicated with exposure longer than 15 min. While some protocols initially suggested culturing the sample for longer time periods [11], extending incubation of the samples to 14 days did not prove to add more positive results compared with a conventional 7-day incubation period [12].

At the end of the 1990s, Tunney et al. developed a sonication protocol for infected hip prostheses, followed by immunofluorescence and PCR (polymerase chain reaction) amplification [13]. Due to a high number of positive results in patients without clinical symptoms or signs of infection (potentially false positives), this method was not incorporated into clinical practice. Later, Trampuz et al. observed that the increased false-positive results associated with sonication could be due to the contamination of cultures by waterborne microorganisms in case of bag leakage and recommended the use of solid containers [14]. Using a protocol designed to avoid this problem, the same authors evaluated sonication in a classic study including a high number of patients with knee and hip prosthetic infections, with extremely good results that improved the sensitivity of periprosthetic tissue culture [15]. In this study, samples were processed in rigid plastic containers and vortexed prior to sonication in a high volume of buffer. Several months later, other studies [16] were published that confirmed the usefulness of sonication, even with different protocols [16, 17]. In the study performed by Esteban et al. [16], the risk of bag contamination was overcome by performing changes of the water in the sonicator before each procedure and by a careful inspection of the bags for leakages [16]. In this study, a concentration step using centrifugation of the sonicate and the use of a broad spectrum of culture media (designed to isolate uncommon organisms) was suggested to improve the sensitivity of the technique. The benefit of centrifugation was confirmed in a recent study, comparing the sensitivity of cultures after membrane filtration (30.3%) versus centrifugation (78.8%) [18]. Other studies using a combined approach with rigid plastic containers and centrifugation confirmed the good results [19].