Orthopedic device-related infections are relatively uncommon (0.6–2.5 % for primary and up to 10–20 % for revision) (Jamsen et al. 2010; Urquhart et al. 2010; Cataldo et al. 2010; Moran et al. 2010; Peel et al. 2011; Shuman et al. 2012). They cause significant morbidity and mortality and an elevated health care expenditure (three times more than first implants).

Surgical risk factors include simultaneous bilateral arthroplasty, a long operative time (more than 2.5 h), and allogenic blood transfusion.

Post-operative risk factors include wound-healing complications such as superficial infection, delayed healing, necrosis, dehiscence, hematoma, bacteremia, prolonged wound drainage, permanence of a urinary catheter, and prolonged hospital stay.

The frequency of infection increases as the number of primary and revision arthroplasties increases (Trampuz and Zimmerli 2006a; Kurd et al. 2010; Laudermilch et al. 2010; Ghanem et al. 2009; Mittal et al. 2007; Estes et al. 2010; Jafari et al. 2010; Mortazavi et al. 2010; Atkins et al. 1998; Zimmerli et al. 2004; Trampuz and Zimmerli 2005).

Almost any microorganism can cause prosthetic joint infections, such as Gram-positive bacteria (accounting for about two-thirds of total), Gram-negative bacteria and polymicrobic flora (accounting for about 10–15 % each), and fungi (rare).

Many classification systems of periprosthetic joint infections have been proposed in the literature. The Centers for Disease Control and Prevention (CDC) defined Surgical Site Infection as an infection of or near the operative site within the first 30 days of surgical procedure or within 1 year with an implant in place (Hidron et al. 2008). Tsukayama’s classification attempted to correlate the time of infection onset with the therapeutic strategy (Tsukayama et al. 1996). His classification, adopted by the Musculoskeletal Infection Society, has been used as a guide for selection of surgical treatment. It defines an early infection as one that occurs within 1 month of index arthroplasty and any infection beyond this point as late. Acute hematogenous infection is also included in this classification system. The Zimmerli/Trampuz classification considers an early infection as one that occurs within 3 months of index surgery. Infections with onset between 3 and 24 months are delayed infections and those occurring >24 months after index arthroplasty are classified as late (Zimmerli et al. 2004; Trampuz and Zimmerli 2005). These classification systems are useful since they provide a description for pathogenesis. It is admitted that early infections may be the result of seeding during surgery, whereas late infections are likely acquired by hematogenous spread. Another classification proposed by Senneville et al. (2009) relies mostly on duration of symptoms and places less emphasis on the timing of index arthroplasty. Based on this classification, acute infection is one with less than 1 month of symptoms and any infection with greater than 1 month of symptoms is considered late. The classification proposed by McPherson considers criteria other than timing, such as host factors and microorganism factors, and looks at periods shorter than 3 weeks (McPherson et al. 2002). Considering the host factors, the McPherson classification quantifies the infection risk factors, making a systemic host grading and a local extremity grading. From the McPherson classification the Musculoskeletal Infection Society scoring system was derived, taking into account the infection type, the systemic host grade and the local extremity grade. The scoring system produces prognostic and therapeutic suggestions regarding the functional results, risks of reimplantation, and risks of limb amputation (McPherson et al. 2002).

Early infections are usually associated with acute onset of pain, effusion, erythema, warmth at the implant site, fever, and increase of inflammatory markers, such as white blood cells, platelet count, C-reactive protein, erythrocyte sedimentation rate, fibrinogen, pro-calcitonin, IL-1, and IL-6 (Greidanus et al. 2007); the formation of a sinus tract with purulent discharge may also occur. Acute infections are usually caused by virulent microorganisms, such as Staphylococcus aureus or, less frequently, Gram-negative bacteria.

Delayed or low-grade infections are most frequently observed and present with insidious signs and symptoms such as implant loosening, persistent joint pain, and non-specific low-grade increase of inflammatory markers. These signs are usually difficult to distinguish from aseptic failure modes. Delayed infections are caused by less aggressive microorganisms such as coagulase-negative staphylococci (CoNS) or streptococci/enterococci. Early and delayed infections are considered acquired from index surgery.

Late infections are predominantly acquired by hematogenous seeding from the skin, respiratory tract, dental apparatus, and urinary or gastrointestinal tract (Trampuz and Widmer 2006).

Other classifications (Coventry 1975) divide the acute postoperative infections into superficial and deep, depending on the infection site.

Superficial infection is limited to the superficial wound layers, above the fascia or capsular layer, and can be treated in most cases by soft tissue debridement with prosthesis retention.

Deep infections below the fascia involving the periprosthetic space are more difficult to treat. In the case of early acute infection starting in the first 3–4 weeks after surgery, an aggressive debridement changing the modular components of the prosthesis may be attempted. In all late chronic deep periprosthetic joint infections, prosthesis removal and exchange (in one or, more commonly, two stages) represents the gold standard.

The key point of all the surgical procedures for fighting periprosthetic joint infection – irrigation and debridement, one stage exchange, or two stage procedures – is the bacteria slime production. Biofilm is the most important factor conditioning the result from irrigation and debridement in acute postoperative infection. Bacterial biofilm production probably starts a few hours after bacterial colonization and prosthesis adhesion but a postoperative period of 3–4 weeks after surgery is allowed for choosing between irrigation and debridement in the case of deep acute periprosthetic infection (Costerton et al. 1999). Further investigations on biofilm are advisable to identify other factors influencing the outcome of the surgical strategies adopted for periprosthetic joint infections.

The management of infections is not completely standardized and the outcome is frequently poor or at least unsatisfactory because of the increasing number of highly virulent and antibiotic-resistant bacteria (Trampuz and Zimmerli 2006 b; Gould et al. 2009; Prokuski 2008).

Considering host factors, patients with infected prostheses usually present with many pre-existing risk factors: obesity, diabetes mellitus, rheumatoid arthritis, other autoimmune diseases, and immunosuppression related to a number of causes (such as previous or active neoplasm, HIV, etc.), tobacco use, previous multiple procedures, and previous infection associated with a prosthetic joint at the same site.

Overweight patients should be counseled to lose weight before surgery. Patients receiving immunosuppressive therapies must be advised to suspend them 2 weeks before surgery and not to resume them for approximately 3 weeks after surgery. Only HIV positive patients should not stop their antiretroviral therapy.

Considering other aspects related to the prevention of prosthetic infections, we must take into account the skin aspects: skin preparation (hair removal), type of skin disinfection, previous scars, adequate blood supply (previous flaps, vein and capillary refilling, arterial local conditions), and presence of pressure sores. The presence of postoperative hematoma is another risk factor because of the more frequent bacterial colonization and the absence of an adequate leukocyte blood supply.

Another aspect which has to be considered is strict glycemic blood level control in diabetic patients. As is well known, high glucose levels worsen leukocyte chemotaxis and blood circulation, reducing the host defenses at the surgical site; moreover, peripheral neuropathy could reduce patient awareness as well as suspicion of infection because of lack of the pain. Furthermore, typical diabetic arterial damage could complicate the postoperative period with acute ischemic limb disease, especially for tourniquet-related knee procedures and for intraoperative leg torsions related to hip procedures. Diabetic patients have a significant increase of ischemic heart and brain events in the postoperative period due to anemia.

Smoking should be strongly discouraged because it increases the infection risk fourfold, probably due to vascular damage. Another possible complication in heavy smokers is COPD exacerbation and an increase in the risk of hospital-acquired pneumonia.

Patients with renal function impairment (especially those on dialysis) and major liver diseases have higher infection risks and drug-related and postoperative side effects, such as allergies, thrombocytopenia and coagulation problems (preexisting and acquired) with the risk of bleeding from the surgical site and hematoma, kidney and liver failure.

Before surgery, the medical team should investigate and take into account the possible treatment of any ongoing source of infection from the teeth and gums, skin, soft tissue, urinary tract, respiratory tract, and gastrointestinal tract.

Prosthetic joint infections are classified by several authors including McPherson et al. (2002), Cierny and DiPasquale (2002), and Cierny et al. (2003).

McPherson developed a staging system for periprosthetic infections, taking into consideration the timing of infection, the overall medical and immune health status of the patient, and the local wound condition.

This staging system considers and integrates three aspects: infection timing type (early <4 postoperative weeks; haematogenous <4 weeks duration; late chronic >4 weeks duration), systemic host grade (A: uncompromised; B: compromised with —one or two compromising factors; C: significantly compromised with more than two compromising factors), and local extremity wound grade (1: uncompromised; 2: compromised with —one or two compromising factors; 3: significantly compromised with more than two compromising factors).