MSIS and EBJIS Criteria | Serum and Synovial Markers | DAIR versus Staged Revision
- MSIS major criteria: two positive cultures or sinus tract
- Synovial WBC greater than 3000 cells per microlitre plus PMN greater than 70 percent supports PJI
- Alpha-defensin has high sensitivity and specificity for PJI diagnosis
- DAIR indicated for acute infection with well-fixed implants and short symptom duration
- Two-stage revision remains gold standard for chronic PJI with high success rates
- “Always aspirate before revision for culture and cell count
- “CRP and ESR together have 90 percent negative predictive value when both normal
- “Leukocyte esterase strip test is rapid bedside synovial marker
- “Biofilm formation explains failure of antibiotics alone in chronic PJI
Two positive periprosthetic cultures with phenotypically identical organisms or a sinus tract communicating with the joint. Either finding alone confirms PJI without need for minor criteria.
Three-tier probability system. Definite PJI requires one major or four minor criteria. Likely PJI requires three minor criteria. Unlikely when fewer than three minor criteria present. Incorporates alpha-defensin and leukocyte esterase explicitly.
WBC count greater than 3000 cells per microlitre with PMN percentage greater than 70 percent in chronic cases. Lower thresholds apply in acute post-operative period (WBC greater than 10000). Always correct for blood contamination.
DAIR within 3 weeks of symptom onset and with well-fixed implants. Chronic infections require implant removal. Delay in diagnosis reduces DAIR success dramatically.
- Diagnosis
- Elevated CRP, positive aspirate, well-fixed implant
- Treatment
- DAIR with modular exchange and IV antibiotics
- Key Pearl
- Success 60-80 percent if early and low virulence
- Diagnosis
- MSIS or EBJIS criteria met, multiple organisms possible
- Treatment
- Two-stage revision with antibiotic spacer
- Key Pearl
- Gold standard, greater than 85 percent eradication
- Diagnosis
- Alpha-defensin positive, borderline cell counts
- Treatment
- One-stage revision if organism known and low virulence
- Key Pearl
- Requires highly sensitive pre-operative workup
FIXEDDAIR Indications and Contraindications
Hook:FIXED implants with short symptom duration allow DAIR success!
Overview and Epidemiology
Periprosthetic joint infection is the most devastating complication after hip and knee arthroplasty. It leads to prolonged hospitalisation, multiple operations, significant morbidity and mortality. Early diagnosis using standardised criteria allows limb salvage with DAIR in acute cases. Chronic PJI almost always requires implant removal. Understanding MSIS and EBJIS criteria, marker thresholds and timing of surgery is essential for examination and safe clinical practice.
- Patient factors: obesity, diabetes, rheumatoid arthritis, smoking, prior infection, immunosuppression
- Surgical factors: prolonged operative time, wound complications, haematoma, simultaneous bilateral surgery
- Implant factors: revision arthroplasty, constrained implants, metal-on-metal bearings
- Post-operative: wound drainage, delayed wound healing, urinary tract infection, bacteraemia
- Incidence: 1-2 percent after primary TKA or THA, up to 5-10 percent after revision
- Timing: 60 percent occur within first 2 years, remainder late haematogenous
- Organisms: coagulase-negative staphylococci and Staphylococcus aureus most common
- Mortality: 2-5 percent at 5 years, higher than many cancers when untreated
Pathophysiology
Periprosthetic joint infection is characterised by bacterial adhesion to implant surfaces and production of extracellular polymeric substance forming a biofilm. Biofilm protects organisms from host immunity and antibiotics, explaining why debridement and implant retention succeeds only in early infection before mature biofilm develops. Low-virulence organisms such as coagulase-negative staphylococci and Cutibacterium acnes produce indolent infections with subtle clinical and laboratory findings. Sonication of explanted components disrupts biofilm and increases culture yield by 30-50 percent in chronic cases.
- Acute PJI
- Early, immature, planktonic dominant
- Chronic PJI
- Mature biofilm with persister cells
- Acute PJI
- Robust inflammatory response, obvious swelling
- Chronic PJI
- Low-grade smouldering inflammation, fibrosis
- Acute PJI
- Markedly raised CRP, synovial WBC very high
- Chronic PJI
- Modest CRP rise, borderline synovial counts
- Acute PJI
- DAIR viable, antibiotics penetrate biofilm
- Chronic PJI
- Implant removal mandatory for eradication
High virulence: Staphylococcus aureus, streptococci, gram-negative bacilli - present acutely with systemic features Low virulence: Coagulase-negative staphylococci, Cutibacterium, enterococci - present late with pain and loosening only Polymicrobial: 10-20 percent of cases, especially in sinus tract or open wounds Culture negative: 5-10 percent despite rigorous sampling, often due to prior antibiotics or fastidious organisms
False negative CRP/ESR: low-virulence organisms, immunosuppression, concurrent antibiotics False positive CRP: recent surgery, inflammatory arthropathy, concurrent infection elsewhere Synovial WBC dilution: bloody tap requires correction formula (subtract 1 WBC per 250 RBCs approximately) Alpha-defensin advantage: less affected by prior antibiotics or inflammatory conditions
Classification and Types
Classification by Timing and Route
- Timing
- Less than 3 weeks after index surgery
- Typical Route
- Intraoperative contamination
- Treatment Preference
- DAIR if implants stable
- Timing
- Greater than 3 weeks but short symptom onset
- Typical Route
- Bacteraemia from distant site
- Treatment Preference
- DAIR if well-fixed implant
- Timing
- Greater than 3 weeks of symptoms
- Typical Route
- Low-grade or late haematogenous
- Treatment Preference
- One or two-stage revision
- Timing
- Months to years
- Typical Route
- Low-virulence organisms
- Treatment Preference
- Two-stage or one-stage if known organism
Timing from symptom onset rather than from index surgery determines surgical strategy. A haematogenous infection presenting 5 years after arthroplasty with 10 days of symptoms is managed as acute.
The 2018 ICM/MSIS Weighted Scoring System
The 2011 MSIS criteria were "all or nothing" for the minor criteria. The 2018 International Consensus Meeting (ICM) update, which the topic cites, replaced this with a weighted point score that examiners increasingly expect candidates to be able to apply. Two positive cultures of the same organism or a sinus tract remain major criteria and are diagnostic on their own; otherwise the points below are summed.
- Threshold
- Over 1 mg/dL
- Points
- 2
- Threshold
- Over 860 ng/mL
- Points
- 2
- Threshold
- Over 30 mm/h
- Points
- 1
- Threshold
- Over 3000 cells per microlitre
- Points
- 3
- Threshold
- Signal-to-cutoff over 1
- Points
- 3
- Threshold
- Positive at ++
- Points
- 3
- Threshold
- Over 80 percent
- Points
- 2
- Threshold
- Over 6.9 mg/L
- Points
- 1
A preoperative aggregate of 6 or more is infected; 2 to 5 is inconclusive and requires intraoperative findings; 1 or less is not infected. Intraoperative findings add further points (positive histology 3, purulence 3, single positive culture 2), and a combined total of 6 or more confirms infection, 4 to 5 is inconclusive, and 3 or less excludes it. This weighted system raised validated sensitivity to 97.7 percent compared with 79.3 percent for the original 2011 MSIS definition.
The two markers most often forgotten are serum D-dimer and synovial CRP - both were introduced in the 2018 weighted score and both appear in the diagnostic infographic. Quote the threshold of 6 points to confirm infection and the 2-to-5 inconclusive band that mandates intraoperative sampling, and remember that a sinus tract or two matching cultures still diagnoses PJI regardless of the score.
CRP ESRMSIS Minor Criteria
Hook:CRP ESR plus synovial thresholds unlock the MSIS minor criteria pathway!
Clinical Assessment
- Onset: acute swelling and pain versus insidious ache
- Systemic features: fever, rigors, night sweats (more common in acute)
- Wound history: delayed healing, prolonged drainage, sinus formation
- Risk factors: recent dental work, urinary infection, skin breach, immunosuppression
- Prior antibiotics: any recent treatment that may mask cultures
- Look: warmth, erythema, swelling, sinus tract, wound dehiscence
- Feel: effusion, warmth, tenderness over joint line
- Move: painful restricted range, instability if loosening present
- Special: check for distant infection sources (teeth, urine, skin)
- Implant stability: assess for loosening signs in chronic cases
Up to 30 percent of chronic PJI cases have normal CRP and ESR, especially low-virulence organisms. Maintain high suspicion in any painful or loose arthroplasty. Always perform aspiration if clinical concern exists regardless of serum markers. A sinus tract or abscess is diagnostic irrespective of laboratory values.
- Distinguishing Features
- Effusion, warmth, sinus, elevated markers
- Investigation Strategy
- Aspiration for cell count, culture, alpha-defensin
- Distinguishing Features
- Gradual pain, no systemic signs, normal markers
- Investigation Strategy
- Serial radiographs, aspiration to exclude infection
- Distinguishing Features
- Mechanical symptoms, giving way, normal markers
- Investigation Strategy
- CT for component position, stress radiographs
- Distinguishing Features
- Sudden pain, trauma history, radiographic fracture
- Investigation Strategy
- CT to assess fracture and implant stability
- Distinguishing Features
- Acute monoarthritis, crystals on microscopy
- Investigation Strategy
- Synovial fluid analysis for crystals and culture
Hip: fluoroscopic or ultrasound-guided anterior or lateral approach. Send synovial fluid for cell count with differential, culture (aerobic, anaerobic, fungal, mycobacterial), alpha-defensin, leukocyte esterase strip, and crystal analysis. Always obtain multiple tissue samples at any subsequent surgery. Blood culture bottles increase yield for fastidious organisms.
Investigations
Diagnostic Workup Sequence
CRP and ESR: CRP greater than 10 mg per L or ESR greater than 30 mm per hour raises suspicion. Combined negative predictive value exceeds 90 percent when both normal. CRP rises faster and normalises quicker than ESR after treatment.
Aspiration mandatory before any revision. Thresholds: chronic WBC greater than 3000 per microlitre with PMN greater than 70 percent; acute post-op WBC greater than 10000. Alpha-defensin sensitivity 95 percent, specificity 95 percent. Leukocyte esterase strip test positive at 2 plus.
Minimum five tissue samples from different sites. Sonication of implants increases sensitivity. Frozen section histology: greater than 5 neutrophils per high-power field (MSIS) or greater than 23 (EBJIS). Gram stain has low sensitivity.
Next-generation sequencing or PCR for culture-negative cases. Metal artefact reduction MRI or CT for collections and bone loss. Nuclear imaging reserved for equivocal cases.
When synovial fluid is blood-stained, correct the WBC count by subtracting approximately one white cell for every 250 red blood cells counted. A grossly bloody tap should prompt repeat aspiration or reliance on alpha-defensin and culture results.
DEFINITEEBJIS Diagnostic Tiers
Hook:Use DEFINITE to remember the full EBJIS probability ladder for PJI.
Culture-Negative PJI
Around 5 to 10 percent of patients who meet the clinical criteria for PJI never grow an organism. This is a recognised and examinable problem because it removes the ability to target antibiotics and to plan a one-stage revision, yet the joint is still infected and must be treated as such.
- Why it lowers yield
- The commonest cause - even short courses suppress growth
- Mitigation
- Withhold antibiotics for at least 2 weeks before sampling whenever it is safe to do so
- Why it lowers yield
- Cutibacterium, fungi and mycobacteria grow slowly or need special media
- Mitigation
- Prolonged incubation (about 14 days for Cutibacterium) and dedicated fungal and mycobacterial cultures
- Why it lowers yield
- Too few samples, swabs rather than tissue, or poor transport
- Mitigation
- Take at least five separate deep tissue samples with fresh instruments; inoculate blood-culture bottles
- Why it lowers yield
- Organisms are adherent, so few are free-floating to culture
- Mitigation
- Sonication of explanted components; molecular testing (16S PCR or next-generation sequencing)
A negative culture in a joint that meets PJI criteria does not mean "no infection". Manage it as PJI - usually a two-stage revision (one-stage is contraindicated without a known organism) - with an empirical broad-spectrum regimen covering staphylococci including MRSA, guided by infectious-diseases input. The priority before any further sampling is to stop antibiotics for two weeks where the patient's condition allows, because this single step recovers many "culture-negative" cases.
Management Algorithm
Acute PJI (Symptom Duration Less Than 3 Weeks) - DAIR Protocol
Indications: well-fixed implants, short symptom duration, no sinus tract, low-virulence or known organism, patient medically optimised.
DAIR Surgical Steps
Position: supine for knee, lateral for hip. Full sterile prep including entire limb. Tourniquet optional for knee.
Reopen previous incision. Perform thorough synovectomy and debridement of all necrotic tissue. Take minimum five tissue samples before antibiotics. Irrigate with 9 litres pulsed lavage.
Mandatory exchange of polyethylene liner or modular head. Inspect backside of components for loosening. Do not retain modular parts.
Close over drains if needed. Commence culture-directed intravenous antibiotics within 24 hours. Transition to oral after 2-6 weeks depending on organism. Total antibiotic duration 3 months.
Success rates 60-80 percent when performed within 3 weeks of symptoms, implants well-fixed, thorough debridement, modular exchange performed, and organism sensitive to biofilm-active antibiotics (rifampicin for staphylococci). Failure rises sharply after 3 weeks or with resistant organisms.
Complications
- Incidence
- 20-40 percent overall
- Risk Factors
- Delayed surgery, resistant organism, no modular exchange
- Management
- Repeat DAIR or conversion to staged revision
- Incidence
- 5-15 percent at 5 years
- Risk Factors
- Immunocompromise, resistant organism, poor soft tissue
- Management
- Repeat two-stage, sometimes amputation
- Incidence
- 10-20 percent
- Risk Factors
- Articulating spacer in non-compliant patient
- Management
- Revision of spacer or early second stage
- Incidence
- Common with prolonged courses
- Risk Factors
- Nephrotoxicity, ototoxicity, C. difficile
- Management
- Therapeutic drug monitoring, shorter courses when possible
- Incidence
- Rare but devastating
- Risk Factors
- Multiple failed revisions, soft tissue destruction
- Management
- Above-knee amputation as salvage
Persistent or recurrent PJI after multiple attempts may necessitate amputation or arthrodesis. Early involvement of plastic surgery for soft tissue coverage and infectious disease specialists for antibiotic planning improves outcomes. Document shared decision-making regarding limb salvage versus amputation.
Outcomes and Prognosis
- Infection Eradication Rate
- 60-80 percent at 2 years
- Functional Outcome
- Retains native joint, good ROM if successful
- Key Determinant
- Timing less than 3 weeks and modular exchange
- Infection Eradication Rate
- 85-95 percent in selected patients
- Functional Outcome
- Single procedure, faster recovery
- Key Determinant
- Known sensitive organism, no sinus
- Infection Eradication Rate
- 85-95 percent at 5 years
- Functional Outcome
- Staged, spacer period impairs function temporarily
- Key Determinant
- Adequate debridement and antibiotic duration
- Infection Eradication Rate
- Near 100 percent infection control
- Functional Outcome
- Major disability, high energy cost
- Key Determinant
- Reserved for failed multiple revisions
Favourable: acute presentation, low-virulence organism, well-fixed implant, compliant patient, no sinus tract. Unfavourable: chronic infection, resistant organism (MRSA, Pseudomonas), immunocompromise, multiple prior surgeries, culture-negative status. Two-stage remains most reliable for chronic PJI across all risk groups.
Guidelines, Registries & Global Practice
- PJI incidence 1-2 percent primary, 5-10 percent revision arthroplasty worldwide
- Most common cause of revision within 2 years in most national registries
- Staphylococci account for 50-60 percent of cases across all regions
- Mortality 2-5 percent at 5 years, comparable to some malignancies
- High-resource: routine alpha-defensin, sonication, next-generation sequencing, articulating spacers
- Limited-resource: reliance on CRP, ESR, basic culture, static spacers, longer antibiotic courses
- Universal principle: thorough debridement and modular exchange matter more than expensive tests
- Surgery: two-stage remains standard globally for chronic PJI; one-stage gaining acceptance in selected cases
- Diagnostic Emphasis
- 2018 weighted criteria, alpha-defensin included
- Acute Treatment
- DAIR for symptoms less than 3 weeks, stable implants
- Chronic Treatment
- Two-stage preferred; one-stage for selected cases
- Diagnostic Emphasis
- Probability tiers, explicit alpha-defensin weighting
- Acute Treatment
- DAIR when criteria met and short duration
- Chronic Treatment
- Two-stage standard; one-stage if organism known
- Diagnostic Emphasis
- Clinical plus laboratory, sonication recommended
- Acute Treatment
- DAIR plus 2-6 weeks IV then oral step-down
- Chronic Treatment
- Two-stage with 4-6 weeks IV antibiotics
- Diagnostic Emphasis
- High index of suspicion, aspiration before revision
- Acute Treatment
- Urgent DAIR pathway for acute presentations
- Chronic Treatment
- Two-stage with multidisciplinary infection team
National joint registries (NJR, AJRR, AOANJRR) consistently identify PJI as the leading cause of early revision. Registry data support two-stage revision for chronic infection with greater than 85 percent infection-free survival at 5-10 years. No randomised trial has demonstrated superiority of one-stage over two-stage in unselected chronic PJI. DAIR outcomes are highly dependent on timing and protocol adherence.
Record in every suspected PJI case:
- Symptom duration in days or weeks from onset
- All serum and synovial marker results with exact values
- Number and results of cultures including sonication
- Implant stability assessment at surgery
- Antibiotic plan and duration with infectious disease input A missed or delayed PJI diagnosis leading to multiple failed revisions is a major source of litigation worldwide. Always document the decision pathway using MSIS or EBJIS criteria.
Controversies & Areas of Uncertainty
Randomised evidence is limited. One-stage offers lower morbidity and cost when strict selection criteria are met (known sensitive organism, no sinus). Two-stage remains safer default for resistant organisms, culture-negative cases or significant bone loss. Surgeon and patient factors often dictate choice.
Most guidelines use 3 weeks from symptom onset, yet some series report acceptable results up to 4-6 weeks. Biofilm maturity is a continuum rather than binary. Decision should integrate organism, implant stability and patient factors rather than rigid time cut-off.
Thresholds and performance in the first 6 weeks after arthroplasty are less well validated than in chronic settings. Some centres still rely primarily on synovial WBC and culture. False positives may occur with recent surgery or haematoma.
No high-quality evidence defines minimum effective duration. Common practice is 2-6 weeks intravenous followed by oral step-down to total 3 months. Shorter courses are being studied in selected low-virulence cases with excellent source control.
MCQ Practice Points
Q: Which finding alone confirms the diagnosis of periprosthetic joint infection according to MSIS criteria? A: A sinus tract communicating with the joint or two positive periprosthetic cultures with identical organisms. These are the two major criteria. Minor criteria (elevated CRP, synovial WBC, alpha-defensin, etc.) are only used when major criteria are not met and require a threshold score.
Q: What is the synovial fluid white cell count threshold suggestive of chronic periprosthetic joint infection? A: Greater than 3000 cells per microlitre with greater than 70 percent neutrophils. In the acute post-operative period the threshold rises to greater than 10000 cells per microlitre. Always interpret in clinical context and correct for bloody taps.
Q: What is the critical time window for considering DAIR in periprosthetic joint infection? A: Less than 3 weeks from symptom onset. Beyond this window biofilm matures, success of DAIR drops below 50 percent, and staged revision becomes the preferred strategy. Symptom duration matters more than time from index surgery.
Q: What are the contraindications to one-stage revision for chronic PJI? A: Sinus tract, polymicrobial infection, resistant organisms (MRSA, VRE), culture-negative status, and significant bone loss requiring structural graft. One-stage requires known sensitive organism, adequate soft tissues and no sinus tract.
Q: What is the clinical utility of synovial alpha-defensin testing in suspected PJI? A: High sensitivity and specificity (approximately 95 percent) even in culture-negative or low-grade infections. Less affected by prior antibiotics than culture. Useful when CRP and cell count are equivocal or when patient has inflammatory arthritis.
Q: How does the EBJIS definition differ from MSIS in the diagnosis of PJI? A: EBJIS uses a three-tier probability system (definite, likely, unlikely) rather than binary. It explicitly weights alpha-defensin and leukocyte esterase and uses a higher histology neutrophil threshold (greater than 23 versus greater than 5 per high-power field). This reduces equivocal cases.
Clinical Imaging
Radiographic and Cross-Sectional Features of Periprosthetic Joint Infection
Plain radiographs may demonstrate periprosthetic lucency, periosteal reaction, or implant loosening but lack sensitivity for early or low-grade PJI. Serial radiographs showing progressive osteolysis are suggestive. CT is useful for assessing bone stock and implant stability prior to revision. MRI with metal artefact reduction sequences can demonstrate soft tissue collections, sinus tracts and marrow oedema. Nuclear medicine scans including leukocyte scintigraphy or FDG-PET have roles when aspiration is inconclusive but are not first-line. Ultrasound can guide aspiration of hip effusions. No imaging modality replaces synovial fluid analysis and culture for definitive diagnosis.
Exam Viva Scenarios
Practise clinical reasoning and management decisions out loud
“A 68-year-old woman develops increasing knee pain, swelling and fever 18 days after primary total knee arthroplasty. CRP is 85 mg per L, ESR 45 mm per hour. Aspiration yields 25000 white cells per microlitre with 85 percent neutrophils. Culture grows methicillin-sensitive Staphylococcus aureus. Radiographs show well-fixed components. What is your diagnosis and management?”
“A 72-year-old man presents 4 years after total hip arthroplasty with a 4-month history of groin pain and a draining sinus over the lateral hip. CRP 25 mg per L, ESR 35 mm per hour. Aspiration grows coagulase-negative staphylococcus in one of three samples. Radiographs show femoral stem subsidence and acetabular osteolysis. How would you manage this patient?”
Diagnostic Criteria
- MSIS major: sinus tract or two identical positive cultures
- Minor criteria: CRP greater than 10, ESR greater than 30, synovial WBC greater than 3000 with PMN greater than 70 percent, alpha-defensin positive, single culture
- EBJIS three-tier: definite (major or four minor), likely (three minor), unlikely (fewer than three)
- Always aspirate before revision; sonication increases yield in chronic cases
Acute versus Chronic Differentiation
- Acute: less than 3 weeks symptoms, high markers, robust inflammation
- Chronic: greater than 3 weeks, modest markers, loosening or sinus possible
- Haematogenous can present acutely even years after index surgery
- Timing from symptom onset determines DAIR eligibility
DAIR Indications and Technique
- Indicated when implants well-fixed, symptoms less than 3 weeks, no sinus
- Mandatory steps: thorough debridement, modular exchange, 9 L lavage
- Send minimum five tissue samples plus sonication fluid
- Post-op: IV antibiotics 2-6 weeks then oral to total 3 months
Revision Strategies
- Two-stage: gold standard for chronic PJI, sinus tract, resistant organisms
- One-stage: selected cases with known sensitive organism, no sinus, good soft tissue
- Spacer: articulating preferred when soft tissue envelope adequate
- Reimplantation only after normalised markers and negative confirmatory aspiration
Key Prognostic Factors
- Favourable: acute, low virulence, stable implant, compliant patient
- Unfavourable: chronic, MRSA or Pseudomonas, sinus, immunocompromise
- DAIR success 60-80 percent when criteria met; two-stage greater than 85 percent
- Failure after multiple attempts may require amputation
Evidence Base and Key Trials
New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society
- Original MSIS consensus definition establishing major and minor criteria for PJI
- Two positive periprosthetic cultures of the same organism, or a sinus tract communicating with the prosthesis, as major criteria sufficient for diagnosis
- Otherwise PJI is diagnosed when FOUR of six minor criteria are met (elevated ESR and CRP, elevated synovial WBC, elevated synovial PMN%, purulence, a single positive culture, and positive histology)
- Standardised diagnostic framework adopted worldwide
The 2018 definition of periprosthetic hip and knee infection: an evidence-based and validated criteria
- Updated MSIS criteria incorporating alpha-defensin and leukocyte esterase
- Weighted scoring system with major and minor criteria
- Validation against 222 revision cases with 97.7 percent sensitivity
- Improved detection of low-grade and culture-negative infections
The EBJIS definition of periprosthetic joint infection
- Three-tier probability system (definite, likely, unlikely) rather than binary
- Explicit inclusion of alpha-defensin, leukocyte esterase and sonication
- Higher neutrophil threshold on histology than MSIS
- Reduces equivocal cases and improves inter-observer reliability
One hundred and twelve infected arthroplasties treated with DAIR (debridement, antibiotics and implant retention): antibiotic duration and outcome
- Retrospective review of 112 PJI cases managed with DAIR over 5 years
- 20 infections (18%) recurred at a mean 2.3 years follow-up (≈82% success); mean antibiotic duration 1.5 years
- Failure was more common after arthroscopic (rather than open) debridement, in previously-revised joints, and with Staphylococcus aureus infection
- Most failures occurred after stopping antibiotics; lengthening antibiotic therapy appeared to postpone rather than prevent failure