TKA Periprosthetic Joint Infection

Concept. DAIR combines aggressive surgical debridement with biofilm-active antibiotics while a stable implant is retained. It is a single-stage retention strategy, but has lower success than 2-stage revision.

Indications. All must be met:

1. Timing: early postoperative (under 4 weeks), or acute haematogenous with symptom duration under 3 weeks
2. Implant stability: well-fixed components, no loosening
3. Tissue quality: intact polyethylene, no gross contamination, viable soft tissue
4. Organism: susceptible to biofilm-penetrating antibiotics (avoid MRSA, fungi, resistant GNR)
5. Host: immunocompetent, with a low KLIC score (under 3)

Antibiotics. Start empirical therapy until cultures return, then change to organism-specific therapy:

• Empirical: vancomycin 15-20 mg/kg IV 12-hourly (target trough 15-20 μg/mL) plus meropenem 1 g IV 8-hourly or ceftazidime 2 g IV 8-hourly
• MSSA: flucloxacillin 2 g IV 6-hourly plus rifampicin 300-450 mg PO 12-hourly; orally, rifampicin 450 mg 12-hourly plus flucloxacillin 1 g 6-hourly
• MRSA: vancomycin (as above) plus rifampicin
• Streptococcus: penicillin G 3 million units IV 4-hourly or ceftriaxone 2 g IV daily; orally, amoxicillin 1 g 8-hourly (rifampicin not required)
• Gram-negative rods: ciprofloxacin 750 mg PO 12-hourly plus rifampicin 450 mg PO 12-hourly (if E. coli, not Pseudomonas)

Duration. IV antibiotics run for 2-6 weeks (debate is ongoing, with some centres giving 2 weeks and others 6), followed by oral suppression for 3-6 months with a biofilm-active, rifampicin-based combination. Rifampicin penetrates biofilm and is active against stationary-phase bacteria, but it is never used as monotherapy, because resistance develops rapidly, within days; always combine it with flucloxacillin, vancomycin or a fluoroquinolone. Two randomised trials in the Evidence Base now bear on these choices: OVIVA found an early switch to oral antibiotics noninferior to IV, and DATIPO found 6 weeks of antibiotics inferior to 12 for PJI overall.

Concept. Two-stage exchange arthroplasty is the gold standard. The infected components are removed and an antibiotic cement spacer placed, the infection is treated systemically, and a new prosthesis is reimplanted once the infection is eradicated. It eradicates infection in 85-90%, but requires two operations and an interval of 6-12 weeks.

Indications.

• Chronic infection (symptom duration over 4 weeks)
• DAIR failure
• Resistant organisms (MRSA, VRE, fungi)
• Loose implants at presentation
• An immunosuppressed host with a high KLIC score

Spacer choice. A static spacer is a block of antibiotic cement with no articulation; an articulating spacer is a knee-shaped mould or a commercial system (PROSTALAC, Spacer-K). Infection eradication is equal (85-90% for both). The articulating spacer gives better range of motion, easier mobilisation and less bone loss than the immobilised static block, but commercial systems cost $3000-5000, while the static spacer is simpler and cheaper.

Many centres internationally prefer articulating spacers in mobile patients for their better function and reduced bone loss during the interval.

Spacer cement. Per 40 g of Palacos R (PMMA):

• vancomycin 4-6 g (4 g standard, 6 g for MRSA or a high bacterial load)
• aminoglycoside 3-4 g tobramycin, or 3.6 g gentamicin
• a maximum of 10% antibiotic by weight, to maintain mechanical strength

Vancomycin gives Gram-positive cover and the aminoglycoside Gram-negative cover and synergy. Together they achieve local concentrations 100 times higher than systemic antibiotics, overcoming the biofilm MIC.

The interval. The standard interval is 6-12 weeks, and there is no evidence that a longer interval improves eradication. Some centres reimplant at 2-4 weeks, for S. aureus with a good response or for virulent organisms with good spacer function; the interval is longer (over 12 weeks) for fungi, resistant organisms and an immunosuppressed host.

Organism-specific IV antibiotics are given for 6 weeks, or 2-4 weeks in some centres for virulent organisms with a rapid response.

Antibiotic holiday. Some centres routinely stop antibiotics for 2 weeks before the second stage (weeks 7-8) to assess for recurrence off antibiotics: the holiday unmasks infection that antibiotics have suppressed, and a CRP rebound suggests persistent infection. Others proceed to the second stage at 6-8 weeks once CRP has normalised, because the holiday delays reimplantation and there is no evidence that it improves eradication rates. Practice is variable, and the trend is no routine holiday when clinical and serological resolution is clear.

Criteria for reimplantation.

1. Clinical resolution: no pain, swelling, erythema or drainage
2. Inflammatory markers normalised: CRP under 10 mg/L (some accept under 20 if falling)
3. Adequate soft tissue coverage: no open wounds, good skin quality
4. Medical optimisation: HbA1c under 7%, adequate nutrition (albumin over 3.5), smoking cessation

These criteria reduce the Stage 2 reinfection risk from 15% to under 5%. Radiographs before Stage 2 assess spacer position and bone stock and plan stems and augments. Aspirating the spacer beforehand is controversial: some centres make it mandatory, others do not aspirate routinely because spacer irritation gives a false-positive rate of 10-20%, and the trend, including in Australia, is no routine aspiration when clinical and serological resolution is clear.

Salvage is reserved for multiple failed revisions, uncontrolled infection or a poor surgical candidate, including the patient with high medical comorbidities. Arthrodesis is rarely feasible in the knee.

Set-up and exposure. Position supine with a tourniquet applied but not inflated during debridement, so that bleeding flushes out bacteria. Use the prior incision, excising the wound edges if contaminated, and make a standard medial parapatellar arthrotomy with the patella everted. Assess the polyethylene and the implant's stability (varus-valgus stress, AP drawer).

Radical synovectomy. Aggressive synovectomy is the key step: incomplete removal leaves a biofilm reservoir and leads to DAIR failure. Remove all synovium from the suprapatellar pouch, the medial and lateral gutters, the posterior compartments (difficult to reach; use curved curettes) and the intercondylar notch, and excise devitalised tissue and haematoma. Curette the bone-implant interface under the femoral component and tibial baseplate to remove interface membrane and granulation tissue. Gross purulence or visibly contaminated cement is a reason to consider removing the cement and converting to 2-stage.

Polyethylene exchange. Unlock the mechanism (it varies by implant system), remove the insert completely, and inspect the tibial baseplate and femoral component for damage. Insert a new liner from a sterile, unopened tray, never a re-sterilised used insert, at the same thickness or adjusted for soft tissue tension, and lock it securely. Exchanging a modular cemented femoral component can be considered, but is controversial.

Irrigation. Change gloves, gown and instruments after debridement and before irrigation, to avoid recontamination. Irrigate with 9-12 litres of normal saline (minimum 6 litres) by pulsed lavage at moderate pressure, which mechanically disrupts biofilm and washes out bacteria; a high-pressure jet drives bacteria into bone. Use no additives: chlorhexidine, betadine and antibiotic solutions are cytotoxic to cartilage and tissue, with no proven benefit.

Closure. Close in layers: the capsule watertight with absorbable suture, the subcutaneous layer to eliminate dead space, and the skin with an absorbable subcuticular suture or staples. Consider a closed suction drain for 24-48 hours, removed when output is under 30 mL per 8 hours; there is some evidence that prolonged drainage (over 7 days) increases reinfection risk, so remove it early. Keep an occlusive dressing on for at least 48 hours.

Success depends on early timing (under 4 weeks), a sensitive organism, complete synovectomy and polyethylene exchange.

Set-up. Position supine; the tourniquet may be inflated for cement removal and then deflated for spacer insertion. Use the prior incision, excising the scar and any sinus tract en bloc.

Explantation. Remove all foreign material, because residual cement is a nidus for biofilm and leads to reinfection:

• the polyethylene insert first
• the femoral component, with osteotomes and a thin flexible saw if cemented, or a slap hammer with offset attachments if uncemented
• the tibial component, with osteotomes from the periphery if cemented, or a slap hammer if uncemented
• all cement (high-speed burr, cement extraction tools, narrow osteotomes), sutures and cerclage wires

Preserve bone stock, avoiding aggressive bone removal unless bone is grossly necrotic or infected.

Sampling and debridement. Take the 5-6 tissue samples described under Investigations, including any purulent or suspicious tissue, for culture and for frozen-section and permanent histology. Then debride radically, removing all synovium (medial, lateral, posterior, intercondylar), the interface membrane (curetting the bone surfaces to bleeding bone), granulation and necrotic tissue and haematoma. Lavage with 9-12 litres of normal saline to reduce the bacterial burden to a minimum before the spacer goes in.

Spacer. Add the antibiotic powders to the cement powder before the liquid monomer and mix thoroughly. For an articulating spacer, hand-mould or use a mould for the femoral component (condyles, trochlea) and the tibial platform, and insert them into bone with gentle pressurisation, not full cementing technique, because the spacer is temporary. Allow full polymerisation (10 minutes) before manipulating the knee; an articulating spacer should then achieve 60-90 degrees of flexion.

View larger

View larger

View larger

Closure. Close in layers over a closed suction drain for 48 hours. A static spacer is protected in a knee immobiliser or a hinged brace locked in extension (or a splint in extension or slight flexion), while an articulating spacer allows immediate range of motion. Weight-bearing is touch only, with a walker, because the spacer is not designed for full load and can fracture.

Timing. Reimplant 6-12 weeks after Stage 1, once CRP has normalised and symptoms have resolved (see the criteria for reimplantation under Management).

Spacer removal. Reopen the same incision, excising the prior scar, and dissect carefully to preserve the extensor mechanism. The spacer may be adherent to bone: free it with gentle osteotomes, avoiding fracture and excessive bone removal. Take 5-6 tissue samples from the synovium and bone surfaces for culture, and a frozen section.

Debridement and bone preparation. Remove residual spacer cement fragments and any granulation tissue, and lavage with 6-9 litres. Bone is often lost to the infection and the spacer interval: classify the defect (Anderson Orthopaedic Research Institute classification, from Type 1 intact metaphyseal bone to Type 3 severe metaphyseal and diaphyseal bone loss), broach or ream to bleeding bone, size for stems if bone is lost, and plan metal blocks or wedges for metaphyseal defects.

Implant selection. Constraint follows bone loss and soft tissue integrity:

• posterior-stabilised, if bone loss is minimal and the collaterals intact (rare after PJI)
• constrained condylar knee, the most common choice for PJI revision, accommodating MCL or LCL deficiency from debridement
• rotating hinge, for severe bone loss, complete collateral deficiency, or extensor mechanism dysfunction or lag

Long cemented stems (100-150 mm) give stability in weakened bone, with offset stems for deformity or bone loss. Augments (metal blocks or wedges) fill femoral or tibial metaphyseal defects and are secured with screws or cement.

Fixation. Cemented fixation is preferred after PJI because it allows local antibiotic delivery, with vancomycin 2 g per 40 g of cement, a lower dose than the spacer to preserve mechanical strength. There is some controversy: antibiotic cement at reimplantation reduces reinfection to 10% from 15%, a marginal benefit. Use full cementing technique with pressurisation, stem centralisers and a 2-3 mm cement mantle, and no tourniquet during cementation, to optimise cement interdigitation.

Soft tissue and closure. Scarring from multiple operations, an attenuated extensor mechanism and skin closure under tension are the challenges. An attenuated or deficient extensor mechanism may be augmented with allograft, or the extensor lag accepted with a long-term brace. Inadequate skin and subcutaneous cover calls for a gastrocnemius flap (the medial head is easier, the lateral head gives better coverage), coordinated with plastic surgery.

Close in layers over a drain for 48 hours, and protect the extensor mechanism in a knee immobiliser or hinged brace for 6 weeks.

Postoperative antibiotics. Practice is variable and controversial. The traditional course is 6 weeks of organism-specific IV antibiotics; many centres give 2-4 weeks of organism-specific antibiotics if intraoperative cultures are negative and 6 weeks if positive, and an emerging practice gives no antibiotics after negative cultures.

The case against antibiotics is that Stage 2 is a "clean" revision if the infection has been eradicated, and prolonged antibiotics may increase C. difficile risk and antibiotic resistance without proven benefit. The case for is a margin of safety in a high-risk host, a resistant organism, or positive intraoperative cultures, even if deemed contamination.

Days 0-2. Continue empirical IV antibiotics until cultures result, then switch to the organism-specific regimen (see Management). Inspect the wound at 48 hours for erythema and drainage. Mobilise on day 0 or 1, sit to stand and walking with a walker, with full weight-bearing as tolerated because the implant is retained and stable, and start passive range of motion on day 1 (CPM or physiotherapy) with the goal of maintaining the preoperative range.

Weeks 1-6. Organism-specific IV antibiotics continue through outpatient parenteral antibiotic therapy (OPAT): a PICC line with home IV administration, or daily infusion clinic visits. Check CRP weekly, inspect the wound weekly for persistent drainage, and assess pain and function. Physiotherapy progresses range and strength towards 0-110 degrees by 6 weeks, and walking from a walker to a cane to independence.

Months 2-6. After the IV course, oral suppression continues for 3-6 months. CRP and clinical assessment are monthly. Activities of daily living return by 3 months, and high-impact sports are avoided lifelong.

Surveillance. Follow up at 3, 6 and 12 months and then yearly, assessing pain, function, range and stability, with CRP annually and yearly radiographs for loosening or progressive lucency (late failure). Teach the patient to report acute pain, swelling or fever promptly, because the risk of acute haematogenous infection is lifelong.

Is DAIR working? CRP should fall by 50% at 2 weeks; if it does not, DAIR is failing. Symptoms should resolve by 4-6 weeks, with a stable range. Persistent pain, a rising CRP or a recurrent effusion indicates failure: proceed to 2-stage.

Days 0-2. Continue IV antibiotics, organism-specific if cultures are available and empirical if pending. Use a knee immobiliser for a static spacer, or a hinged brace unlocked for range of motion with an articulating spacer. Weight-bearing is touch only, with a walker, and the drain comes out at 48 hours.

Weeks 1-6. IV antibiotics continue through OPAT. CRP is checked weekly and should normalise to under 10 by 6 weeks, with weekly clinical assessment for pain, drainage or erythema. Weight-bearing stays limited to touch; an articulating spacer allows range-of-motion exercises, with the goal of maintaining 60-90 degrees. Spacer complications (see Complications) may need a return to theatre.

The antibiotic holiday and the pre-reimplantation assessment are covered under Management.

Weeks 0-6. A knee immobiliser or hinged brace is locked for 6 weeks to protect the extensor mechanism and, with a constrained implant, to prevent dislocation; it is locked in extension for walking and unlocked for range-of-motion exercises under physiotherapy supervision. Weight-bearing starts as touch only with a walker, progresses to partial (50%) by 2 weeks and to as tolerated by 6 weeks. Passive range of motion (CPM or physiotherapy) aims for 0-90 degrees by 6 weeks, accepting less where scarring limits motion.

Any antibiotics given run IV for 2-6 weeks and then stop, with no long-term suppression after a successful 2-stage. CRP at 2 and 6 weeks should remain low; a rise suggests reinfection and needs urgent investigation.

Weeks 7-12. Wean the immobiliser at 6 weeks if the extensor mechanism is intact and the wound healed, and progress to full weight-bearing, then a cane, then independence. Quadriceps strengthening is critical, because extensor lag is common after 2-stage, and hydrotherapy is beneficial. Active range of motion aims for 0-100 degrees by 12 weeks (accept 90, since scar tissue limits gains), and radiographs at 6 and 12 weeks assess fixation, lucency and alignment.

Months 3-12. Full recovery takes 6-12 months, longer than an uncomplicated primary TKA. Manipulation under anaesthesia at 6-12 weeks is considered if range is under 70 degrees and interfering with function. Return is to low-impact activities (walking, swimming, cycling), with high-impact sports avoided lifelong.

Surveillance is at 3, 6 and 12 months, then yearly, with CRP annually and radiographs yearly for 5 years.

Antibiotic prophylaxis. Timing is critical:

• Standard: cefazolin 2 g IV (3 g if weight over 120 kg) within 60 minutes before incision (ideally 30-60 minutes), with a second intraoperative dose if surgery lasts over 4 hours or blood loss exceeds 1500 mL
• Non-anaphylactic penicillin allergy: cefazolin is still acceptable (10% cross-reactivity, but severe reactions are rare)
• History of anaphylaxis: vancomycin 15 mg/kg IV (infused over 1 hour, started 90-120 minutes before incision) plus a single dose of gentamicin 5 mg/kg IV
• Known MRSA colonisation: vancomycin 15 mg/kg IV plus cefazolin 2 g IV, dual coverage being superior to vancomycin alone

A single dose is standard. Prophylaxis prolonged beyond 24 hours increases C. difficile and resistance without reducing PJI.

Skin preparation. Chlorhexidine-alcohol (2% chlorhexidine in 70% isopropyl alcohol) is superior to povidone-iodine, reducing SSI by 40% against iodine (RCT evidence). Apply it twice, with a 3-minute drying time each, over a wide field (hip to ankle for the knee), and keep chlorhexidine away from mucous membranes and eyes (ototoxicity, and neurotoxicity if it enters the joint accidentally). Clip hair only if it interferes with the incision, and immediately before surgery rather than the night before, which allows bacterial regrowth; do not shave, because microabrasions increase infection.

Theatre environment. Laminar flow reduces the airborne bacterial count by 90% compared with conventional ventilation and is standard in many arthroplasty units. Each door opening disrupts laminar flow for 5 minutes, so minimise openings and limit personnel to essential staff. Body exhaust suits are controversial: there is some evidence of reduced infection, but they are uncomfortable and expensive and not universally adopted.

Maintain normothermia (core temperature over 36°C), because hypothermia increases infection risk by 30% through vasoconstriction and impaired neutrophil function.

Surgical technique. Each additional 30 minutes of operating time increases PJI risk by 15%, so work efficiently without rushing, and handle tissue gently, avoiding excessive retraction and cautery, to preserve the blood supply to the skin edges. Meticulous haemostasis reduces haematoma, a substrate for bacterial growth; tranexamic acid reduces blood loss and haematoma formation and may reduce PJI risk (emerging evidence).

Irrigate with 3-6 litres of pulsed lavage before closure, using normal saline only, since chlorhexidine, betadine and antibiotics are cytotoxic. Close in layers, avoiding dead space; absorbable subcuticular sutures and staples have similar infection rates. Antibiotic cement for the implant can be considered in high-risk patients (revision, immunosuppressed).

Drains. Proponents argue that a drain reduces haematoma, a substrate for infection; opponents that it is a foreign body and a route for retrograde bacterial entry, with no proven benefit, and meta-analysis shows no difference in PJI with or without drains. Practice is variable, and many surgeons use a closed suction drain for 24-48 hours, removed when output is under 30 mL per 8 hours. The consensus is that a drain, if used, comes out early (under 48 hours), and prolonged drainage (over 7 days) increases PJI risk 3-fold.

Wound. Leave an occlusive dressing (hydrocolloid or film) for 48 hours minimum, and do not remove it to "check the wound", which disrupts the seal and lets bacteria in; after 48 hours, inspect daily for erythema and drainage. Persistent drainage over 7 days carries a high PJI risk, and a return to theatre for irrigation and closure should be considered. A superficial dehiscence is not packed: return to theatre for debridement and reclosure under sterile conditions.

Glycaemic control. Target a glucose under 10 mmol/L (180 mg/dL) postoperatively; hyperglycaemia over 12 mmol/L in the first 48 hours increases PJI risk 2-fold. Diabetic patients receive a sliding-scale insulin protocol postoperatively, with an endocrinology consult if control is poor.

Thromboprophylaxis. Aspirin 100 mg daily or rivaroxaban 10 mg daily for 14 days (AAOS guidelines), with mechanical prophylaxis (TED stockings, foot pumps); the same regimen applies after DAIR. There is no evidence that anticoagulation increases PJI, but haematoma from inadequate haemostasis or excessive anticoagulation may increase infection risk, so use prophylactic doses only.

Mobilisation. Physiotherapy on day 0 or day 1 (sit to stand, walking with a walker) reduces DVT, improves respiratory function and reduces complications. There is no evidence that early mobilisation increases wound complications or PJI if the wound is sealed.

PJI is the leading or second leading indication for revision TKA in virtually every national registry, and its absolute burden is rising worldwide as primary arthroplasty volume grows.

The US economic-burden data are from Kurtz et al, J Arthroplasty 2012 DOI. Across high-income registries the primary TKA infection rate clusters around 1-2% and revision around 3-5%, but absolute numbers are climbing everywhere with arthroplasty demand.

Major guideline bodies broadly agree on diagnostic and surgical principles but differ in emphasis, diagnostic thresholds and the strength of evidence behind individual recommendations.

The IDSA management framework is from Osmon et al, Clin Infect Dis 2013 DOI and the 2018 weighted definition from Parvizi et al, J Arthroplasty 2018 DOI.

Where they differ: AAOS and the IDSA no longer recommend routine antibiotic prophylaxis before dental procedures for joint replacement patients, whereas several other bodies and individual surgeons still advise it for high-risk patients (shared decision-making). Diagnostic cut-offs (e.g. synovial PMN over 70% vs over 80%, synovial CRP, alpha-defensin availability) vary between definitions, and the threshold for offering single-stage exchange is more permissive in parts of Europe than in North America.

Antibiotic Route and Duration: Landmark RCTs

The route and duration of antibiotics referenced throughout this chapter (the "IV 2-6 weeks", "oral suppression" and post-Stage-2 duration debates) are now informed by two landmark randomised trials. These resolve the PJI-specific decisions raised here; general antibiotic pharmacology and resistance are developed in the orthopaedic-antibiotic-therapy topic.