Septic Arthritis

Septic arthritis is a bacterial infection within a joint space — a true orthopedic and medical emergency, not merely another cause of a painful joint. Once bacteria establish themselves in synovial fluid, the combination of bacterial toxins, host inflammatory cytokines, and neutrophil-derived proteolytic enzymes can destroy articular cartilage within 24 to 48 hours, and that destruction is irreversible — cartilage has essentially no capacity to regenerate.^[1][2] Every hour spent debating the diagnosis is an hour the joint is actively being destroyed, which is why this chapter is organized around a single, non-negotiable safety message: any acutely painful, swollen, warm joint with restricted range of motion must be treated as septic arthritis until proven otherwise, with urgent joint aspiration — not a fever curve, not a white cell count, not a wait to see how the patient trends — as the diagnostic step that actually rules the condition in or out.^[3]

This chapter covers the classic causative organisms by age and risk group, empiric antibiotic selection pending culture results, the essential, non-optional role of urgent surgical joint drainage alongside (never instead of) antibiotics, prosthetic joint infection as a distinct and more complex management scenario, and the important, easily missed diagnostic nuance that gout and pseudogout can look identical to septic arthritis at the bedside and can even coexist with it in the same joint.

Pathophysiology and routes of infection

The synovial membrane lining a joint has a rich blood supply and, unlike most other tissues, no limiting basement membrane separating it from the vascular space — a combination that makes the joint an unusually easy target for hematogenous seeding during transient or sustained bacteremia, the dominant route by which most native joints become infected.^[2][1] Once organisms are established in the synovium, bacterial proteases and toxins combine with the host's own neutrophil-driven inflammatory response to degrade cartilage matrix — a process that begins within hours, not days.^[1] Two other routes matter clinically: direct inoculation, from a penetrating injury, an intra-articular injection, arthroscopy, or open joint surgery; and contiguous spread from an adjacent focus of osteomyelitis, particularly in young children, whose metaphyseal blood vessels cross the growth plate into the joint space.

Septic arthritis of the ankle in a 3-month-old infant. (Hagino T, Wako M, Ochiai S, CC BY 2.5, Wikimedia Commons)

Clinical presentation

The classic presentation is acute-onset monoarticular pain, swelling, warmth, and erythema with markedly restricted active and passive range of motion — the restriction of passive range of motion in particular is a genuinely useful discriminating feature, since periarticular processes such as bursitis or cellulitis typically still allow passive movement of the joint itself. The knee is the single most commonly affected joint, followed by the hip (especially in children), shoulder, wrist, and ankle; roughly 10-20% of cases involve more than one joint, and polyarticular involvement is associated with higher morbidity and mortality and should raise suspicion for bacteremia, endocarditis, or a disseminated process.^[2]

The distinct presentation of disseminated gonococcal infection

Disseminated gonococcal infection (DGI) typically begins with a prodrome of migratory polyarthralgias and tenosynovitis — often affecting the wrists, fingers, knees, and ankles in a fleeting, moving pattern — before evolving along one of two overlapping patterns: a purulent, often monoarticular or oligoarticular arthritis that looks like classic septic arthritis, or the "arthritis-dermatitis" syndrome — the triad of tenosynovitis, dermatitis (a small number of painless or minimally tender vesiculopustular or hemorrhagic lesions, typically on the distal extremities near joints), and polyarthralgia, without frank purulent arthritis.^[1][5] This distinction genuinely matters at the bedside: in the arthritis-dermatitis pattern, joint fluid culture is frequently negative even when the diagnosis is correct, and blood culture, cervical/urethral/pharyngeal/rectal nucleic acid amplification testing (NAAT), and culture of a skin lesion are all more likely to be revealing than the joint aspirate alone — a genuinely high-yield point, since a negative joint fluid culture is sometimes wrongly taken as evidence against the diagnosis.^[1] Risk factors specific to dissemination (as opposed to acquiring uncomplicated gonorrhea in the first place) include female sex, menstruation, pregnancy, and terminal complement pathway deficiency, including iatrogenic complement inhibition with an agent such as eculizumab.^[6]

• Bacteremia from any source (skin/soft-tissue infection, an IV catheter, a dental procedure, endocarditis) — the dominant route of infection for most native joints.^[2]
• Pre-existing joint disease, especially rheumatoid arthritis or another inflammatory arthropathy, or a joint already damaged by osteoarthritis or prior injury — an already-inflamed, hyperemic joint is more susceptible to hematogenous seeding, and pre-existing joint disease also makes a new infection harder to recognize against a background of chronic pain and swelling.
• A prosthetic joint — a distinct risk category with its own diagnostic thresholds and management approach (see Special Populations).
• Age over about 80, or age under 2-3 years.
• Diabetes mellitus, chronic corticosteroid use, or another cause of immunosuppression (HIV, malignancy, chemotherapy, biologic/immunosuppressive therapy).
• Chronic kidney disease, particularly on hemodialysis, and cirrhosis.
• Injection drug use.
• Recent intra-articular corticosteroid injection, arthroscopy, or joint surgery.
• A skin or soft tissue infection, ulcer, or breach overlying or near a joint.
• Sexual risk factors for gonococcal arthritis: young age, multiple partners, unprotected intercourse; menstruation, pregnancy, and terminal complement deficiency specifically raise the risk of dissemination once infection occurs, rather than the risk of acquiring gonorrhea itself.^[6]
• Sickle cell disease.
• Malignancy.

• Any acute monoarticular or oligoarticular joint pain and swelling with restricted range of motion, especially with fever, chills, or another sign of systemic infection.
• Any new or worsening joint pain, swelling, wound drainage, or systemic symptoms in a patient with a prosthetic joint — this needs urgent orthopedic assessment using prosthetic-joint-specific diagnostic thresholds, not the native-joint pathway (see Special Populations).
• Hip pain or refusal to bear weight in a child — the hip is a deep joint that is difficult to examine and follow serially, carries the highest risk of avascular necrosis of the femoral head if infection is missed or delayed, and should prompt urgent orthopedic involvement rather than a period of outpatient observation.
• Polyarticular involvement, or any sign of hemodynamic instability, sepsis, or bacteremia — raises the stakes and the need for a broader source-control workup (e.g., for endocarditis) alongside joint-specific management.
• Any immunocompromised patient (diabetes, corticosteroid use, HIV, malignancy, dialysis, cirrhosis) with new joint symptoms — presentation can be blunted and atypical, so the threshold to aspirate should be lower, not higher, in this group.
• New joint or bone pain in a patient with sickle cell disease — do not assume this is simply a vaso-occlusive crisis without actively considering septic arthritis or osteomyelitis, particularly if the pain pattern, joint findings, or fever are not typical for the patient's usual crises.
• Suspected disseminated gonococcal infection — hospitalize for initial management and involve infectious diseases; this reflects the systemic nature of the process, not just the joint findings.^[6]

History and examination

Characterize onset (acute versus gradual), the number of joints involved, and screen specifically for the risk factors and red flags above: recent bacteremia or skin infection, a prosthetic joint, immunosuppression, injection drug use, and sexual history/risk factors relevant to gonococcal disease. On examination, assess for warmth, erythema, effusion, and — the most discriminating single physical finding — marked restriction of both active and passive range of motion, since passive motion is often preserved in periarticular processes such as bursitis, cellulitis, or tendinopathy.

Laboratory testing

Obtain a complete blood count, ESR, and CRP, but interpret them cautiously: each has only weak individual discriminatory power (positive likelihood ratios of roughly 1.3-1.6 for an elevated serum white cell count, ESR, or CRP), and a normal result does not meaningfully lower the probability of septic arthritis once clinical suspicion is present.^[3] Obtain two sets of blood cultures before starting antibiotics whenever possible — roughly half of adult, non-gonococcal cases have positive blood cultures, and a positive result both supports the diagnosis and can identify the organism when joint fluid culture is unrevealing.^[2][1]

Arthrocentesis — the diagnostic cornerstone

Joint aspiration is the single most important diagnostic step, and should be performed before the first dose of antibiotics whenever feasible without meaningfully delaying treatment of a septic or hemodynamically unstable patient — prior antibiotic exposure reduces synovial fluid culture yield.^[1] Send synovial fluid for cell count and differential; Gram stain; aerobic and anaerobic culture (with NAAT for N. gonorrhoeae/C. trachomatis where gonococcal disease is a possibility); and crystal analysis under polarized microscopy — every time, regardless of how confidently the clinical picture "looks like" gout or pseudogout (see the crystal-arthropathy discussion below).

Gram stain identifies the organism in roughly half of confirmed staphylococcal/streptococcal cases, but sensitivity is considerably lower for gram-negative organisms and for gonococcus specifically — a negative Gram stain never rules out septic arthritis on its own.^[1][2]

Imaging

Plain radiographs are a reasonable baseline (they can show a joint effusion, pre-existing joint damage, or an adjacent bony abnormality) but are usually normal early in the disease course and should never delay arthrocentesis. Point-of-care ultrasound is useful to confirm an effusion and can guide aspiration of a joint that is difficult to access clinically, notably the hip. MRI is the most sensitive imaging study for early joint infection and for detecting adjacent osteomyelitis, and is particularly useful for the pediatric hip and for suspected prosthetic joint infection — but obtaining it should never come before, or meaningfully delay, urgent aspiration and orthopedic involvement when clinical suspicion is high.

A pediatric-specific tool: the Kocher criteria for the irritable hip

In a child with an acutely irritable hip, the Kocher clinical prediction algorithm combines four independent predictors — inability to bear weight, ESR ≥40 mm/hr, fever, and serum WBC >12,000/µL — to estimate the probability of septic arthritis versus the far more common and benign transient synovitis.^[8] A subsequent validation cohort found the rule performs less well prospectively than at its original derivation, and predicted probabilities should be interpreted with that in mind:

^[9]

Use this table as one input into clinical judgment, not as a stand-alone decision rule — it was derived primarily from a cohort dominated by S. aureus infection, and it performs less reliably in children under about 4 years old, where Kingella kingae is proportionally more important and can present more subtly than the classic picture the rule was built around.^[8][4] A low predicted probability should lower, not eliminate, clinical concern in a child who otherwise looks unwell.

• Confirm the diagnosis urgently with arthrocentesis, without letting the workup delay definitive treatment once infection is strongly suspected.
• Achieve source control with urgent surgical drainage of the infected joint — this is not optional, and antibiotics alone do not substitute for it.^[1]
• Start empiric intravenous antibiotics covering the likely organism(s) by age and risk factors immediately after cultures are obtained, without waiting for results.
• De-escalate to targeted, narrow-spectrum therapy once an organism and susceptibilities are known.
• Preserve joint function and prevent irreversible cartilage destruction, adjacent osteomyelitis, and systemic sepsis.
• Recognize and manage prosthetic joint infection as its own, distinct problem — different diagnostic thresholds, and a surgical strategy decision (retain vs. remove the prosthesis) that drives the entire treatment plan.
• Actively look for and, where present, treat coexisting crystal arthropathy — without ever letting a positive crystal result substitute for excluding infection.
• For gonococcal disease, treat the systemic/disseminated process, not just the joint, and complete the associated STI-management bundle: partner notification, empiric partner treatment, and screening for other STIs.

Neonates and young children

Neonates cefotaxime has historically been preferred over ceftriaxone in this age group because ceftriaxone can displace bilirubin from albumin, raising kernicterus risk in a hyperbilirubinemic neonate or one receiving calcium-containing IV fluids; where cefotaxime is unavailable, ceftriaxone can be considered case-by-case in a low-risk neonate without hyperbilirubinemia, but this should be a deliberate, informed substitution rather than a default one.^[11] Pair with an antistaphylococcal agent (e.g., vancomycin if MRSA risk, or cloxacillin) based on local epidemiology and severity. Children under about 2-3 years cefazolin provides reliable coverage of both MSSA and Kingella kingae, the two dominant pathogens in this age group, and is the empiric agent used in current Canadian pediatric guidance; broaden to include MRSA coverage (e.g., add vancomycin or clindamycin, depending on local resistance patterns) if risk factors or a concerning Gram stain suggest it.^[4]

De-escalation and duration

Once an organism and susceptibilities are available, narrow to targeted therapy — for example, cloxacillin or cefazolin for confirmed MSSA, continued vancomycin (or an alternative such as daptomycin or linezolid) for MRSA, and a susceptibility-directed agent for a gram-negative isolate. Total duration for native-joint, non-gonococcal septic arthritis has traditionally been about 3-4 weeks (longer, up to 6 weeks, for more severe S. aureus disease or a slow clinical response), but the evidence base has been shifting toward shorter courses when source control has genuinely been adequate: a randomized non-inferiority trial in non-immunocompromised adults found that 2 weeks of antibiotic therapy after thorough surgical drainage was non-inferior to 4 weeks (remission ~99% vs. 97%).^[12] Individualize the actual duration to the organism, adequacy of source control, and clinical trajectory rather than applying one fixed number to every patient, and favour an early switch from intravenous to a highly bioavailable oral agent once the patient is clinically improving, afebrile, and able to tolerate oral intake — an approach with strong trial-level support in bone and joint infection generally (see the Prosthetic Joint Infection discussion under Special Populations).^[13] For gonococcal arthritis/arthritis-dermatitis syndrome specifically, switch to an oral agent guided by susceptibility testing once there has been substantial clinical improvement for 24-48 hours, for a total treatment course of more than 7 days.^[6]

A note on Canadian guidance for gonococcal disease

The Public Health Agency of Canada's Canadian Guidelines on Sexually Transmitted Infections identify disseminated gonococcal infection as a recognized complication of gonorrhea but do not publish their own dose-and-duration regimen for it — instead, PHAC directs that complicated gonococcal infections, including disseminated disease, be managed by or in consultation with an infectious diseases specialist or an experienced colleague, with hospitalization for initial management.^[14] The CDC regimen above is the most detailed, currently in-use clinical guidance available and is presented here for that reason, to be applied in the context of local/current infectious diseases consultation.

• Urgent surgical joint drainage (arthrotomy or arthroscopic washout) is the central non-pharmacologic intervention in this chapter — it is not an adjunct to antibiotics, it is co-equal source control, and delaying it while "seeing how the antibiotics do" is a recognized cause of preventable joint destruction.^[1][10]
• Repeat aspiration or repeat surgical washout if the joint reaccumulates fluid or the clinical response is inadequate — a single washout is not always sufficient, particularly for the knee.
• Splinting/immobilization in the acute phase for comfort, followed by early, progressive mobilization and physiotherapy once the infection is controlled — prolonged immobilization risks stiffness and contracture, especially at the knee and hip.
• Adequate analgesia, including short-term opioid therapy where needed — the pain of an acutely septic joint, and of the surgical drainage itself, is often severe, and under-treating it is a real, avoidable source of patient distress.
• Source control beyond the joint itself when bacteremia is present: remove or exchange an infected indwelling catheter, and evaluate for endocarditis (e.g., with echocardiography) in S. aureus bacteremia or a patient with a prosthetic heart valve, since the joint may be one manifestation of a broader bloodstream infection.
• Partner notification, empiric partner treatment, and a full STI screening bundle (chlamydia, syphilis, HIV) for confirmed or strongly suspected gonococcal arthritis — treat and notify partners from the preceding 60 days, consistent with standard Canadian STI-management practice, and report confirmed gonococcal infection to local public health.^[14]
• Multidisciplinary coordination (orthopedic surgery, infectious diseases, and, for a prosthetic joint, the original implanting surgeon where possible) for any prosthetic joint infection, since the surgical strategy decision drives everything else (see Special Populations).

Pediatrics

Apply the same non-negotiable safety principle as in adults: a normal temperature or normal white cell count does not exclude septic arthritis in a child. In a child with an acutely irritable hip, the Kocher criteria (see Assessment & Diagnosis) can help estimate probability, but should never substitute for urgent orthopedic involvement when clinical suspicion is high, and perform less reliably in children under about 4 years old, where Kingella kingae is proportionally more important.^[8][4] Current Canadian pediatric guidance from the Canadian Paediatric Society recommends empiric cefazolin 100-150 mg/kg/day divided every 6-8 hours, covering the two dominant pediatric pathogens (MSSA and Kingella kingae); broaden coverage if the Gram stain, local epidemiology, or clinical severity suggests MRSA or another organism.^[4] A meaningful, real shift in duration has occurred in recent years: a total antibiotic duration of 3-4 weeks (transitioning to oral therapy once the child is clinically well and tolerating oral intake) is now typical for most uncomplicated cases with a good response — down from a previously standard 6 weeks — while hip septic arthritis specifically is still generally treated for a longer 4-6 weeks, given its higher stakes and historically higher complication rate.^[4] Distinguish septic arthritis from the far more common and benign transient synovitis using the clinical picture as a whole (transient synovitis is associated with a lower-grade fever, gradual improvement over days, and a CRP typically under 20 mg/L) rather than any single test.^[4]

Prosthetic joint infection

Infection involving a prosthetic joint is a genuinely distinct clinical problem, not simply "septic arthritis in someone who happens to have a joint replacement" — the diagnostic thresholds are different, the causative organisms skew toward lower-virulence skin flora that would be dismissed as contaminants in a native joint, and the treatment decision is fundamentally a surgical one about whether the prosthesis itself can be saved.

Diagnosis: different thresholds, and a sinus tract is diagnostic on its own

A sinus tract communicating with the joint or visible hardware is diagnostic of prosthetic joint infection by itself, without further testing.^[15] Otherwise, synovial fluid WBC thresholds are considerably lower than in a native joint and depend on timing since implantation: a widely used, evidence-based definition uses roughly 3,000 cells/µL with more than 80% neutrophils as the threshold for chronic infection (generally more than about 90 days after implantation), compared with a substantially higher threshold, on the order of 10,000 cells/µL, in the acute postoperative period, when the normal inflammatory response to surgery itself elevates synovial cell counts.^[16] Applying the native-joint threshold (tens of thousands of cells/µL) to a prosthetic joint will miss real infections. Obtain ESR and CRP as adjunctive, not stand-alone, tests; synovial fluid aspiration for cell count, differential, and culture — ideally with a meaningful antibiotic-free window beforehand if the patient's condition allows it, to maximize culture yield; and, at the time of surgery, multiple (at least three, ideally five to six) intraoperative tissue cultures, since a single positive culture in this setting is genuinely harder to interpret than in a native joint and can reflect contamination rather than true infection.^[15]

Surgical strategy — the decision that drives everything else

Debridement, antibiotics, and implant retention (DAIR) is a reasonable strategy specifically for a well-fixed prosthesis with no sinus tract, presenting within roughly 30 days of implantation or within about 3 weeks of symptom onset.^[15] Outside that window — a loose or poorly fixed prosthesis, a sinus tract, a late/chronic presentation, or a difficult-to-treat organism — prosthesis removal, as either a one-stage or, more commonly in North American practice, a two-stage exchange (removal, an antibiotic-impregnated spacer, a course of antibiotics, then reimplantation once infection is controlled), is generally required.^[15] This is the genuinely more complex part of prosthetic joint infection management that distinguishes it from native-joint septic arthritis: the surgical decision is not simply "drain and treat," but a structured choice between salvaging and sacrificing the implant, made jointly by orthopedic surgery and infectious diseases.

An antibiotic-loaded acrylic bone cement hip spacer, placed after removal of an infected acetabular component as the first stage of a two-stage exchange; a definitive prosthesis is implanted once the infection has resolved. (Hellerhoff, CC BY-SA 4.0, Wikimedia Commons)

Rifampin — a genuinely distinct pharmacologic point

For staphylococcal prosthetic joint infection managed with DAIR (implant retained), adding rifampin in combination with a companion oral agent (commonly a fluoroquinolone) targets bacteria within the biofilm that antibiotics alone struggle to reach — rifampin is never used as monotherapy, given how readily resistance emerges. Total duration is 3 months for a hip and 6 months for a knee prosthesis, started once surgical wound drainage has resolved.^[15]

Oral therapy is increasingly appropriate — a current Canadian position

A 2025 joint statement from AMMI Canada and the Canadian Orthopaedic Association (developed with the Canadian Arthroplasty Society) supports using highly bioavailable oral antibiotics — not only as a late step-down, but from relatively early in the course — for adult bone and joint infections, including prosthetic joint infection, once: adequate source control has been achieved surgically; the causative organism is confirmed and susceptible to an appropriate oral agent; there is no other reason IV therapy is specifically required; the patient has a functioning GI tract and can be expected to adhere to a 6-12 week course; and this is done in collaboration with infectious diseases.^[13] This builds directly on the landmark OVIVA trial, which established oral therapy's non-inferiority to IV therapy for bone and joint infection more broadly, and on the 2013 IDSA guideline framework above.^[13]

Gonococcal arthritis and disseminated gonococcal infection

See Recognize the Condition and Choose Initial Treatment for the distinct clinical presentation and treatment regimen. A patient with recurrent or unusually severe disseminated gonococcal infection, or a family history suggestive of it, warrants consideration of an underlying terminal complement pathway deficiency — a rare but recognized predisposing condition — and referral for further workup.^[6] Complete the associated STI-management bundle for every confirmed case: partner notification and empiric treatment, screening for chlamydia/syphilis/HIV, and public health reporting.^[14]

Immunocompromised patients and people who inject drugs

Both presentation and laboratory findings can be blunted in an immunocompromised patient, so maintain a lower threshold to aspirate rather than a higher one, and broaden empiric gram-negative coverage (e.g., ceftazidime or piperacillin-tazobactam rather than ceftriaxone) when injection drug use, neutropenia, or another specific pseudomonas risk factor is present. In people who inject drugs, actively examine and consider imaging of atypical, easily overlooked joints — the sternoclavicular, sacroiliac, and pubic symphysis joints — since infection at these sites is both disproportionately represented in this population and more likely to be missed on a routine joint examination.

Sickle cell disease

Salmonella species are a classic and disproportionately represented cause of septic arthritis and osteomyelitis in sickle cell disease, alongside S. aureus. The genuinely important, high-stakes trap in this population is diagnostic: new bone or joint pain is easy to attribute to a routine vaso-occlusive crisis, and doing so without actively considering infection risks a missed or delayed diagnosis in a population already prone to both problems. A pain presentation that is atypical for the patient's usual crises, localizes to a single joint with objective swelling/warmth/effusion, or is accompanied by fever should prompt the same urgent aspiration-first approach used for any other patient.

Pregnancy and breastfeeding

Do not let pregnancy delay urgent aspiration or surgical drainage — the risks of untreated septic arthritis to both mother and fetus substantially outweigh the risks of appropriate treatment. Preferred empiric agents vancomycin, ceftriaxone, and cefazolin are all generally considered compatible with pregnancy and are reasonable choices for empiric or targeted therapy. Avoid where possible fluoroquinolones (e.g., levofloxacin) given animal data on cartilage effects and generally limited human safety data, and doxycycline/tetracyclines beyond about the first trimester, given the risk of fetal tooth discoloration and bone effects — if a fluoroquinolone-rifampin combination is genuinely needed for a prosthetic joint infection in pregnancy, this is a specialist infectious diseases and obstetric decision, not a routine one. Rifampin itself crosses the placenta but is used when a serious maternal infection genuinely requires it; involve infectious diseases and obstetrics jointly. Vancomycin, the penicillins/cephalosporins, and clindamycin are generally considered compatible with breastfeeding at standard doses; avoid fluoroquinolones and, in the neonatal period particularly, sulfonamide-containing agents given theoretical kernicterus risk in a jaundiced or unwell infant.

Track clinical response daily in the acute phase: pain, swelling, range of motion, and fever trend, alongside serial CRP (which tends to trend down meaningfully faster than ESR and is generally the more useful marker for following response) — expect a clear clinical and biochemical improvement within the first several days once source control and appropriate antibiotics are both in place; a plateau or worsening picture should prompt reassessment rather than simply extending the current plan (see Treatment Failure/Next Step).

• Confirm the total antibiotic duration and route match the clinical scenario: roughly 3-4 weeks for native-joint, non-gonococcal septic arthritis in most adults (individualized — as short as 2 weeks after genuinely thorough surgical drainage in a non-immunocompromised adult with a rapid, clear response^[12]); 3-4 weeks (4-6 weeks for the hip) in children^[4]; more than 7 days total for gonococcal arthritis/arthritis-dermatitis syndrome, with an oral switch guided by susceptibility 24-48 hours after substantial improvement^[6]; and 3 months (hip) or 6 months (knee) of rifampin-combination therapy for staphylococcal prosthetic joint infection managed with DAIR.^[15]
• Confirm final culture and susceptibility results are reviewed and used to de-escalate or, if needed, broaden therapy — do not leave a patient on empiric broad-spectrum coverage by default once results are available.
• Reassess ESR/CRP before any planned reimplantation surgery in a two-stage prosthetic joint exchange, as part of confirming the infection has genuinely been controlled before proceeding.^[15]
• Arrange physiotherapy and structured rehabilitation once the acute infection is controlled — restoring range of motion and strength is a real, ongoing part of recovery, not an afterthought.
• For gonococcal disease: confirm partner notification and empiric partner treatment occurred, and screen for reinfection at around 6 months post-treatment given the shared sexual-network risk.^[14]
• Arrange durable orthopedic follow-up after any prosthetic joint infection, including monitoring for relapse and, in some retained-prosthesis cases, a discussion of long-term suppressive oral antibiotic therapy when the infection cannot be fully eradicated.

1. No clear improvement in pain, swelling, or inflammatory markers within about 48-72 hours of appropriate antibiotics and adequate initial surgical drainage: reassess with repeat arthrocentesis and imaging to look for inadequately drained fluid, a loculated collection, adjacent osteomyelitis, or a resistant/unexpected organism — and reconsider the antibiotic regimen in light of any new culture information.
2. Any evidence that source control was incomplete (a reaccumulating effusion, persistent purulence, an inaccessible or under-drained joint — especially the hip): arrange repeat urgent surgical washout rather than simply extending or broadening antibiotics.
3. Persistent bacteremia, a new murmur, or another sign concerning for endocarditis, particularly with S. aureus: pursue echocardiography and a broader source-control workup — the joint may be one manifestation of a wider bloodstream infection.
4. A prosthetic joint infection not responding to DAIR: reconsider one- or two-stage exchange rather than repeating debridement indefinitely — persisting with implant retention beyond what the clinical picture supports risks chronic, difficult-to-eradicate infection.^[15]
5. Crystal-positive synovial fluid in a patient whose clinical picture still does not fully fit simple gout/pseudogout (disproportionate pain, systemic symptoms, an atypical joint, or risk factors for infection): do not stop pursuing infection — continue empiric antibiotics pending full culture incubation if suspicion remains genuine, since the two can coexist.^[7]
6. Culture-negative septic arthritis with high clinical suspicion: hold cultures for extended incubation (fastidious organisms such as Kingella kingae can take longer to grow), consider NAAT-based testing where available, and reconsider fungal or mycobacterial causes in a patient with a compatible exposure history or risk factors.
7. Recurrent infection in the same joint: evaluate for retained infected hardware, chronic osteomyelitis, or an underlying immunodeficiency — including, for recurrent gonococcal dissemination, a terminal complement pathway deficiency — rather than simply repeating the same treatment course.

• Treating a normal temperature or a normal serum white cell count as evidence against septic arthritis — neither reliably rules it out.^[1][3]
• Waiting for imaging to be arranged and read before involving orthopedics or performing arthrocentesis when clinical suspicion is genuinely high — every hour of delay is an hour of active cartilage destruction.
• Starting antibiotics before obtaining synovial fluid and blood cultures when this is feasible, which measurably reduces the chance of identifying the causative organism.^[1]
• Treating confirmed or strongly suspected septic arthritis with antibiotics alone, without arranging urgent surgical drainage — antibiotics do not substitute for source control in this condition.^[1][10]
• Dismissing a negative synovial fluid culture as evidence against gonococcal arthritis — the arthritis-dermatitis pattern in particular is frequently joint-culture-negative, and blood, skin-lesion, and mucosal NAAT/culture are often more revealing.^[1]
• Assuming a positive crystal analysis (gout or pseudogout) rules out coexisting infection — send Gram stain and culture on every joint aspirate regardless of crystal findings.^[7]
• Applying native-joint synovial WBC thresholds to a prosthetic joint — the genuinely applicable threshold is much lower, and further depends on how recently the joint was implanted.^[16]
• Giving an empiric corticosteroid injection, or assuming a "flare" diagnosis, in a joint where infection has not actually been excluded.
• In a child with hip pain, delaying orthopedic involvement while awaiting labs or imaging when clinical suspicion is high — the hip's poor accessibility and avascular-necrosis risk make this a poor place to "wait and see."
• Attributing new bone or joint pain in a patient with sickle cell disease automatically to a vaso-occlusive crisis without actively considering septic arthritis or osteomyelitis.
• Starting rifampin as monotherapy, or before surgical wound drainage has resolved, for prosthetic joint infection.

How to use this table: Doses are adult doses and require patient-specific confirmation (renal/hepatic function, weight, allergy history). Pediatric dosing is weight-based per the current product monograph unless stated. Drug pricing is intentionally omitted. Select any drug name to open its complete Pharmasist Drug Database page.

Empiric Intravenous Therapy — Native-Joint Septic Arthritis, Adults

Obtain synovial fluid and blood cultures before the first dose whenever feasible. Cover MRSA and likely gram-negative organisms pending results; broaden for pseudomonas risk (injection drug use, immunocompromise, hospital-acquired infection).

Targeted Therapy Once an Organism and Susceptibilities Are Known

De-escalate to the narrowest effective agent once culture results are available.

Gonococcal Arthritis / Disseminated Gonococcal Infection

Hospitalize for initial management and involve infectious diseases. Switch to an oral agent guided by susceptibility testing 24-48 hours after substantial improvement, for a total course of more than 7 days.

Prosthetic Joint Infection — Biofilm-Directed Adjunct Therapy

For staphylococcal PJI managed with DAIR (implant retained). Start rifampin only once surgical wound drainage has resolved, and always in combination.

Pediatric Empiric Therapy — Children Under About 2-3 Years

Covers the two dominant pediatric pathogens (MSSA and Kingella kingae); broaden for MRSA risk if indicated. Neonates: cefotaxime is preferred over ceftriaxone when hyperbilirubinemia is present, given ceftriaxone's bilirubin-displacement risk.

Crystal Arthropathy Flare — Once Septic Arthritis Is Addressed or Confidently Excluded

Never given empirically for a presumed inflammatory monoarthritis before septic arthritis has been excluded — corticosteroids and, to a lesser extent, NSAIDs can mask an unrecognized infection.