Modern-day total joint replacement was pioneered in the 1960s for treatment of arthritis [1]. With advances in surgical technique, coupled with an ageing population, this procedure continues to grow in popularity throughout the world. US data predict that the number of hip or knee replacements will increase by over 40% from 2015 to 2020, a prediction echoed in other international studies [2–4]. Evidence supports both the clinical and cost-effectiveness of this procedure [5].

Throughout the early attempts and refinement of the procedure, infection of the prosthesis has remained a constant foe [1]. However, capturing the exact burden and risk of infection in this cohort has proven elusive. Determination of the true incidence of infection is hampered by a number of factors including the lack of uniform definition of prosthetic joint infection, lack of robust post-discharge surveillance for infection, and imprecise methods for capturing infections in registries [6–8]. Overall, the accepted incidence of infection is thought to be approximately 1–3%, that is, infection of the prosthesis is an uncommon complication [9]. These infections, however, have a profound effect on patient morbidity and mortality, in addition to being a significant economic burden. Of concern, the risk of infection is increasing and is predicted to exceed 6.5% by 2030 [10].

The presence of the biofilm impacts on all aspects of diagnosis, management, and prevention of prosthetic joint infection [11]. The causative pathogens associated with prosthetic joint infection are generally skin-associated microorganisms, with a predominance of Gram-positive organisms such as coagulase-negative staphylococci . There is, however, an increasing incidence of infections due to Gram-negative bacteria observed in a number of studies [12]. Given the changing global ecology, with the emergence of antimicrobial resistance, particularly in Gram-negative bacilli, this observation is concerning and the overall impact of these organisms on elective joint replacement surgery is unknown [12].

In addition to the role of biofilm in the development of infection, patient-related factors also play a role. Given that this surgery is undertaken in older patients, comorbid diseases, such as diabetes mellitus and obesity, influence surgical outcomes. These are potentially modifiable factors; however, there are limited data to demonstrate that optimisation of these factors influences outcomes. These factors are the current focus of a number of studies.

Wound complications , including haematoma and wound ooze, have also been implicated in the development of prosthetic joint infection. Whether these wound complications are distinct entities from infection or are, in fact, early manifestations of an (as yet) undiagnosed infection remains an issue of conjecture. Furthermore, while early intervention for prolonged postoperative wound drainage is advocated, whether this intervention reduces the incidence of prosthetic joint infection is unknown [13].

In recent years, the publication of classification schemes and diagnostic criteria for prosthetic joint infection has provided a consistent framework to guide detection and management of infection. Recognition of the clinical, diagnostic, and management differences between acute and chronic infections is important. Clinical features may guide diagnosis but are imperfect; up to 15% of patients coming to revision arthroplasty with the preoperative presumptive diagnosis of aseptic failure have positive intraoperative cultures [14–16]. Development of diagnostic criteria, such as those published by the Infectious Diseases Society of America or by the Musculoskeletal Infection Society, not only aids diagnostic acumen but also to enables comparison of different tests [17, 18].

Correct differentiation between ‘septic’ and ‘aseptic’ failure remains the holy grail of research into prosthetic joint infection diagnosis. Conventional microbiology laboratory culture techniques have a low sensitivity [16]. This is further compounded by the fact that many of the common pathogens of prosthetic joint infection, such as coagulase-negative staphylococci , are also the main cause of specimen contamination.

Microbiological culture is the cornerstone of diagnosis for confirmation of infection and to assess antimicrobial susceptibility to guide treatment and preventative strategies. The likelihood of detecting infection, and differentiating infection from contamination, increases with increasing number of specimens provided [16]. In addition, collection of periprosthetic tissue samples, rather than swabs, increases yield and minimises contamination [19]. Current guidelines recommend that a minimum of three tissue specimens be obtained for microbiological culture; however, these guidelines further suggest five to six specimens is the ideal number to optimise diagnosis [16, 17].

Recent research has focused on efforts to optimise or augment laboratory techniques. This includes the application of sonication techniques , which act to release the bacteria from the biofilm attached to prostheses, thereby increasing the microbiological yield [20]. In addition, adaptation of laboratory culture methods, such as inoculation of synovial fluid and periprosthetic tissue specimens into blood culture bottles, has increased the sensitivity and accuracy of diagnosis [21–23]. Of note, when blood culture bottles for periprosthetic tissue specimen culture are used, the number of specimens required for accurate diagnosis is reduced compared with conventional techniques [24, 25].

There has also been reinvigorated interest in conventional markers such as C-reactive protein (CRP) , erythrocyte sedimentation rate , and synovial fluid white cell parameters to provide more robust data on the utility of these tests. This research has provided increased clarity around diagnostic thresholds for these markers in acute and chronic settings and according to different joints [18].

In addition, new markers have been investigated, including leucocyte esterase , synovial CRP, and alpha-defensin . Initial research has suggested that these markers may be associated with improved sensitivity with preserved specificity. In particular, the early reports of alpha-defensin were particularly impressive with reported sensitivity of 97%, far in excess of conventional diagnostic tests [26]. While promising, more extensive validation of alpha-defensin is required, particularly in acute or chronic infection and in different joints. The role of these new synovial markers will be determined in the coming years. Radiology, while commonly performed, remains of uncertain utility.

Overall, given the diagnostic challenges and uncertainties, there should be a low threshold to consider infection. All patients undergoing revision surgery should have assessment of serum inflammatory markers, synovial fluid analysis, and multiple tissue samples for culture, in addition to histological examination to maximise the detection or exclusion of prosthetic joint infection.

As noted, the management of prosthetic joint infection is determined by the chronicity of infection and by the type and susceptibilities of the infecting organism(s). Neither surgery nor antibiotics administered in isolation lead to successful treatment outcomes. Therefore, a multidisciplinary approach to management with the input of both surgical and medical expertise is critical to the success of prosthetic joint infection management. In addition, understanding the effect of different antimicrobials on biofilm-associated microorganisms influences treatment outcomes.

The goals of treatment must be clearly defined to guide management decision; in particular, determining whether the goal is cure versus suppression of the infection is of import. The three broad treatment approaches are curative with prosthesis retention, curative with prosthesis removal , and infection suppression without removal of the prosthesis. Ultimately, the ideal management approach should ensure eradication of infection with preservation of a pain-free, functional joint [11].

Curative approaches with prosthetic retention can be considered in acute infections, without clinical or radiological evidence of prosthesis loosening and where the causative microorganisms are susceptible to antimicrobials with a good anti-biofilm profile [17]. Patients with chronic infection, prosthesis loosening, or infections due to microorganisms where there are no suitable anti-biofilm antimicrobial options available, and where the intent is curative, should undergo removal of the prosthesis as part of one- or two-step exchange procedures, or of joint arthrodesis [17]. Long-term suppressive antimicrobials, with retention of the prosthesis, are a reasonable management strategy in patients with significant comorbidities that would preclude more aggressive, curative approaches [17].

There have been a number of studies investigating the anti-biofilm properties of different antimicrobials both in laboratory and clinical settings. Overall, there are increasing data to support the use of rifampicin in infections with Gram-positive organisms where the intent is cure with retention of the prosthesis. The main limitation of using rifampicin is its low threshold for generation of resistance; therefore this antibiotic cannot be administered alone and must be given with a companion antimicrobial. The weight of evidence favours combination with levofloxacin as the treatment of choice for staphylococcal prosthetic joint infection. There are limited data to guide antibiotic choices for implant retention with other Gram-positive organisms. Evidence to support antimicrobial strategies for Gram-negative infections is likewise limited; however, ciprofloxacin is the accepted agent for first-line treatment [17].