Septic Arthritis

Septic arthritis is a type of arthritis caused by a bacterial infection entering a joint via an open wound or from the bloodstream. Symptoms of septic arthritis appear suddenly and may include swelling, tenderness, and fever. If pus builds up, the joint may be permanently damaged.

Fluid is taken from the joint and is analysed to determine the presence of infection (see aspiration), and pus may be drained to help relieve pain. Initially, treatment is with intravenous antibiotic drugs, followed by oral antibiotics for several weeks or months after that. Septic arthritis is also known as pyogenic arthritis.

Read more: Septic arthritis - diagnosis and management - technical

Septic arthritis in detail - technical


Acute pyogenic arthritis may be primary (by haematogenous spread) or secondary (to trauma, surgery, or arthrocentesis). Organisms that cause primary septic arthritis are usually aggressive pathogens: Staphylococcus aureus (most commonly), streptococci, salmonella species (especially in African children) and Gram-negative organisms (neonates and older people). The causes of secondary and chronic septic arthritis are more diverse, including lower-grade pathogens from skin, mycobacteria, and fungi.

Clinical presentation—patients with acute pyogenic arthritis typically present with fever and an acutely painful joint that is swollen (effusion), warm to the touch, tender on palpation, and painful—frequently exquisitely so—on active or passive movement. The host response appears to reduce the risk of bacteraemia and death, but at the cost of joint damage. If not fatal through septicaemia, untreated septic arthritis generally causes joint destruction or fusion, sometimes with sinus formation and persistent infection. Chronic native joint septic arthritis presents in more indolent fashion with a mono- or polyarthropathy.

Diagnosis, management and prognosis—the diagnosis of pyogenic arthritis is established by isolation of a recognized pathogen from samples of synovium or synovial fluid obtained through biopsy or aspiration. After obtaining blood cultures and (whenever possible) synovial fluid, acute pyogenic arthritis should be treated promptly with intravenous antibiotics active against aerobic Gram-positive cocci and, where appropriate, Gram-negative organisms. Two or three weeks of antibiotic treatment is usually given in uncomplicated cases. Urgent consultation with an orthopaedic surgeon is recommended: arthroscopic washout is usually recommended, although some cases can be managed by joint aspiration once or twice daily until clinical response is evident. If acute native joint infection is treated promptly, the prognosis is good. Many patients make a complete recovery, but joint damage is highly likely when the diagnosis is made late. Outcomes are less favourable in prosthetic joint infection.


Pyogenic arthritis, which may be acute or chronic, describes infection and resulting inflammation in a joint, native or prosthetic. It should not be confused with postinfectious (reactive) arthritis (discussed here: reactive arthritis). As with other musculoskeletal infections (see: osteomyelitis), failings in diagnosis or management may have long-term functional consequences, and clinicians should therefore know when to consider the diagnosis and obtain expert help.

Aetiology, pathogenesis, and pathophysiology

Acute pyogenic arthritis may be primary (by haematogenous spread) or secondary (to trauma, surgery, or arthrocentesis). Organisms that cause primary septic arthritis are usually aggressive pathogens capable of causing a bacteraemia, seeding a joint, and multiplying within it—hence they are also common causes of septicaemia. Staphylococcus aureus dominates in most circumstances, but streptococci are also important, salmonella species have been found to be important causes in African children, and Gram-negative organisms are more common in neonates and older people. The causes of secondary and chronic septic arthritis are more diverse. In prosthetic joints they also include lower-grade pathogens from skin. In native joints, mycobacteria and fungi are seen as causes of chronic septic arthritis alongside all the causes of acute infection when this has not been cured. Table 1 shows the common pathogens involved in both pyogenic arthritis and osteomyelitis.

In primary septic arthritis, organisms exit the bloodstream to access the joint. In the case of S. aureus, invasion of endothelial cells can occur through interactions between bacterial fibronectin-binding proteins and cell surface-associated fibro-nectin. This triggers integrin-dependent uptake of bacteria and may be a key first step in seeding of sites of metastatic infection during bacteraemia.

S. aureus releases a number of toxins and proteases thought to affect host defences. It also expresses numerous cell-wall-associated adhesins that mediate attachment of bacteria to the matrix proteins associated with cell surfaces, cartilage, and bone. Animal models demonstrate that T-cell dependent inflammation plays a central role in damage to articular cartilage following an acute inflammatory response. In these models immunomodulation (e.g. with corticosteroids) can substantially reduce arthritis, but at the expense of host survival if antibiotic therapy is not also given. Thus the host response appears to reduce the risk of bacteraemia and death, but at the cost of joint damage. If not fatal through septicaemia, untreated septic arthritis generally causes joint destruction or fusion, sometimes with sinus formation and persistent infection through the involvement of bone and dead cartilage as foci.

In a prosthetic joint, the presence of foreign material impairs local antibacterial defences. As with dead bone in chronic osteomyelitis, the bacteria adhere to inanimate material (in this case, plastic or metal) to produce a community called a biofilm. In this state they become relatively resistant to the action of antibiotics and to phagocytic killing. Ineffective but chronic inflammation causes pain and triggers bone loss at the interface with the infected implant or its associated bone cement, with subsequent mechanical loosening, pain, and failure of the joint.


The incidence of pyogenic arthritis has been estimated at 7 in 100 000 in northern industrialized nations: the incidence is highest in children and older people, and more common in males, but there are large worldwide variations, with a chance of 1 in 1000 that a child in Malawi will develop the condition in the first 5 years of life. The increased incidence in older people probably reflects a higher prevalence of potential sources of bacteraemia, such as urinary tract infection, skin ulceration, pneumonia, and hospitalization with intravenous and/or urinary catheterization. Infection complicates some 0.5 to 2% of total joint replacements, but the true prevalence of prosthetic joint infection is unknown, and there is evidence of variation in rates between centres.

Prevention and control

There are no proven means of preventing primary pyogenic arthritis. Secondary cases can be prevented by meticulous attention to infection control measures whenever a joint is aspirated or operated on, and by thorough cleaning, debridement, and antibiotic treatment when a joint is contaminated through trauma. Prevention of prosthetic joint infection is optimized by fastidious sterile technique, ultrafiltered laminar air flow, and prophylactic antibiotics. The evidence that secondary prophylaxis is necessary or beneficial for patients with prosthetic joints in situ who are undergoing medical or dental procedures is not of high quality; some expert bodies recommend prophylaxis should accompany dental procedures occurring within 1 to 2 years of implantation of a prosthetic joint.

Clinical features

Patients typically present with fever and an acutely painful joint. The joints adjoining the long bones are most commonly affected (knee, hip, shoulder, elbow, wrist, and ankle). There may be bacteraemia (in some series in up to 70% of cases), causing prostration, vomiting, or hypotension. Infants localize pain poorly and commonly present refusing to use the affected limb. Adults may be unable to localize pain if a sternoclavicular, acromioclavicular, sternocostal, or manubriosternal joint is involved. Infection in these locations often presents as atypical chest wall pain. Sacroiliac joint infection causes buttock or low back pain and may mimic hip or spine pathology.

Clinical examination reveals a joint that is swollen, warm to the touch, tender on palpation, and painful, frequently exquisitely so, on active or passive movement. This presentation should be considered as infection until proven otherwise. To minimize pain, the patient will often nurse the joint in a neutral position. A joint effusion is usually present and this may be accompanied by synovitis, depending on the duration of the history. Erythema is not usually prominent and if present, may signal the presence of bursitis, cellulitis, or a more superficial abscess.

The pain of acute pyogenic arthritis generally resolves within the first 1 to 2 weeks of successful treatment, but stiffness and swelling usually persist for very much longer.

Chronic native joint septic arthritis presents in more indolent fashion with a mono- or polyarthropathy and a variable degree of joint destruction. There may be systemic features or evidence of involvement of other organ systems (for example, but not always, in tuberculosis). There may be strong epidemiological clues to exposure to unusual pathogens through the travel or occupational history, or there may be a history suggestive of an earlier acute septic arthritis. Patients with peripheral neuropathy may present with an obvious chronic septic arthritis underlying a neuropathic ulcer that has penetrated to the joint.

Prosthetic joint infection may present as an acute wound infection, a periarticular abscess, an acute arthritis, or loosening of the implant (as progressive and chronic pain). A sinus discharging in or near the operative scar represents infection of the prosthesis until proved otherwise. Loosening of an implant within a few years of primary surgery should raise the suspicion of infection unless there is an obvious mechanical problem.

Differential diagnosis

Septic arthritis must be distinguished from other acute or chronic monoarthropathies: notably gout, pyrophosphate arthropathy, and haemarthrosis. Rheumatoid and reactive arthritis can initially present with involvement of only a single joint. Infection is most commonly monoarticular, but multiple joints can be involved. Polyarticular infection may be mistaken for a flare in the underlying disease in a patient with known inflammatory arthritis. Trauma to the joint involving vegetable material, such as thorns, may introduce infection, but the organic matter can also lead to a very aggressive sterile inflammatory arthritis. In the case of a prosthetic joint, the differential is from superficial wound infection, haemarthrosis, periprosthetic fracture or dislocation, and aseptic loosening.

Clinical investigation

The criterion for diagnosis of pyogenic arthritis is the isolation of a recognized pathogen from samples of synovium or synovial fluid obtained through biopsy or aspiration. Infected synovial fluid is generally turbid or purulent. It should be sent for Gram stain, semiquantitative or quantitative white cell count, examination under polarized light for pyrophosphate or uric acid crystals, and culture. If tuberculosis, brucellosis, or fungi are suspected, the laboratory should be advised so that the sample can be appropriately processed. If Neisseria gonorrhoeae is suspected, urethral, endocervical, throat, and rectal swabs should also be obtained for microscopy and culture.

Blood cultures should always be obtained in suspected acute infection. The white cell count, C-reactive protein, and ESR are usually raised, but can also be elevated during flares of inflammatory arthritis or acute crystal arthropathy. Their value is probably greatest in following the response to treatment. Measurement of serum uric acid may be elevated in gout, but cannot be used to establish or refute this diagnosis. Serological tests may be of retrospective value in diagnosing Borrelia burgdorferi (Lyme disease), and brucella. Detection of microbial nucleic acid in joints remains a research technique, but this is likely to change as technologies become increasingly robust.

Plain radiographs can show fracture, effusion, chondrocalcinosis, bone destruction, or loss of joint space. An effusion may be seen acutely, but bony changes appear over at least 10 to 14 days. If seen at presentation they either indicate chronic infection or represent underlying arthropathy. A radiograph at presentation provides a useful baseline for subsequent comparisons. The role of CT or MRI is to demonstrate or exclude surgical disease in the joint or neighbouring bone. Ultrasonography may assist in this and in distinguishing between effusion and synovitis, allowing diagnostic samples to be obtained more reliably.

Criteria for diagnosis

Consensus criteria for diagnosing infection have not been agreed. The isolation of microbes from joint fluid, capsule, or synovium is pathognomonic of infection as long as contamination of the specimen(s) can be ruled out. Cellular pathology is a valuable confirmatory test, with synovial fluid containing polymorphs and the synovium showing an acute inflammatory response with a fibrinous exudate on its surface, which may be ulcerated. A chronic synovitis may develop, with a lymphocytic and mononuclear infiltrate. Tissue from infected prosthetic joints shows a polymorph infiltrate accompanied by chronic inflammatory changes representing a reaction to foreign materials. Where low-grade or fastidious pathogens are suspected, multiple independent specimens can increase diagnostic yield and help to distinguish contamination of samples from true infection.


Acute septic arthritis poses a threat to the joint and is a musculoskeletal (orthopaedic or rheumatological) emergency. Treatment should generally be in an inpatient hospital setting. Patients with suspected chronic septic arthritis may initially be managed on an outpatient basis, but most eventually require surgical intervention.

After obtaining blood cultures and (whenever possible) synovial fluid, acute pyogenic arthritis should be treated promptly with intravenous antibiotics active against aerobic Gram-positive cocci and, where appropriate, Gram-negative organisms. Suitable regimens include a high-dose semisynthetic antistaphylococcal penicillin (flucloxacillin, dicloxacillin, or nafcillin), with or without an aminoglycoside; a β-lactam–β-lactamase combination (amoxicillin–clavulanate or ampicillin–sulbactam); or cefuroxime (or another antistaphylococcal cephalosporin). Patients allergic to β-lactams, or with risk factors for methicillin-resistant S. aureus (MRSA), should receive vancomycin, usually with an aminoglycoside, until culture results are obtained. The specific choice of empirical regimen should be determined with microbiological input based on local sensitivity patterns. Definitive treatment regimens should be based on culture results.

Urgent consultation with an orthopaedic surgeon is advised. Arthroscopic washout has largely replaced arthrotomy, reducing morbidity. Surgery can sometimes be avoided altogether by aspiration once or twice daily until clinical response is evident, but may still be needed if there is clinical deterioration or failure to settle within 5 days. If the joint is accessible this strategy can be applied to children or adults, particularly when anaesthesia is thought to carry high risks, e.g. in the patient with multiple comorbidities. Whether delaying surgery in this way gives worse outcomes than immediate washout is unknown. There is broad consensus on the need for prompt surgery on the hip, where the capsular vessels are reflected up the neck of the femur and are vulnerable to thrombosis, which causes avascular necrosis of the femoral head. If this occurs, joint destruction, with or without chronic infection of the dead bone, is inevitable. In the shoulder, despite the similar anatomy, it is not clear if arthrotomy or washout is superior to aspiration.

The optimal duration and mode of administration of antibiotics is unknown. In uncomplicated infection, 2 to 3 weeks is probably adequate, depending on the pathogen (2 weeks for streptococci, 3 weeks for S. aureus and aerobic Gram-negative rods). In children it is possible to convert to oral therapy within 48 to 72 h of defervescence, provided that there has been a rapid clinical response. This strategy requires the organism to be sensitive to a reliably bioavailable oral antibiotic, the parents or carers to understand clearly the importance of adhering to the antibiotic regimen, and the clinician to monitor clinical progress carefully. Some authorities treat adults with an oral regimen provided that similar criteria are met, whereas others prefer intravenous antibiotics when there has been accompanying bacteraemia or where there are concerns about bony involvement, absorption of oral antibiotic, or adherence to the treatment regimen. Many patients are suitable for outpatient intravenous antibiotic therapy, provided this is properly organized and supervised. Longer courses of treatment are indicated if there have been complications, slow resolution, or suspected involvement of underlying bone.

Chronic septic arthritis has generally led to joint destruction, with death of cartilage and involvement of underlying bone, by the time the patient presents. The condition is thus usually a form of chronic osteomyelitis (see osteomyelitis). Surgical debridement is generally necessary to control infection, and staged arthrodesis or joint replacement—after an infection-free interval— the most frequent reconstructive strategy. Surgery to remove or to attempt to salvage the joint is also the rule in prosthetic joint infection, although for a few patients a decision not to operate may be the option with the least associated morbidity. Antibiotic treatment is usually prolonged in chronic or prosthetic joint infection, and may be long term where a decision is made to attempt to suppress, rather than cure, the infection.


If diagnosed and treated promptly, the prognosis of acute native joint infection is good, with many patients making a complete recovery. Conversely, joint damage is highly likely when the diagnosis is made late, for which reason clinical guidelines have been produced to aid appropriate early management. Infection in young children may lead to disturbance of the growth plate around the infected joint, causing deformity. Mortality is low in uncomplicated septic arthritis, higher when it is complicated by S. aureus bacteraemia (up to 20%), and highest in polyarticular disease (50%). Recurrence is uncommon and generally indicates a persisting surgical focus in the joint.

Outcomes are less favourable in prosthetic joint infection. Salvage of the original infected prosthesis can be achieved in 30 to 70% of cases where the prosthesis is retained. Infection can be eradicated in up to 90% of cases with revision surgery, but with very much poorer results expected when revision surgery for infection is itself complicated by further infection. This and the need for expert surgical and microbiological input, as well as the considerable comorbidity such patients often have, makes the management of infected prosthetic joints a formidable challenge best undertaken by, or with the support of, a specialist multidisciplinary team in a major centre.

Chronic infections have significant effects on quality of life through pain, poor function, and general debility. This is most prolonged and severe in prosthetic joint infection, which may culminate in amputation if sufficiently persistent or recurrent to cause major bone and soft tissue loss. For these reasons, while patients with acute infection may experience some psychosocial issues in adjusting to their abrupt change in functional ability and to uncertainties over the extent of functional recovery, patients with chronic infections usually have more obvious difficulties, especially when there is a nosocomial cause for their problems.

Areas of uncertainty or controversy

Numerous aspects of treatment await clarification in prospective studies, including the relative merits of arthrotomy, arthroscopic washout, or repeated aspiration; the optimal duration and route of administration of antibiotics; and the possible benefit of adjunctive steroids to modify destructive inflammatory responses in native joint infection.

Likely future developments

The dominance of staphylococci as a cause of septic arthritis makes likely a continuing increase in the prevalence of infections due to methicillin-resistant organisms. In some centres this is already driving changes in empirical treatment regimens and in the choices for definitive therapy. Molecular methods of diagnosis will yield a greater understanding of the range of causative pathogens, particularly in previously culture-negative cases and in prosthetic joint infections, and are likely to become standardized tools. It remains to be seen whether trends in rising antimicrobial resistance and rising numbers of elderly patients at risk of native and prosthetic joint infection will be balanced or overcome by improvements in surgical technique, coordinated programmes to reduce healthcare-associated infections, and investigational compounds that might reduce the ability of bacteria to adhere to implants. Septic arthritis is expected to remain a diagnostic and therapeutic challenge for some time to come.

Further reading

Read more: Septic arthritis - diagnosis and management - technical


Berendt AR (1999). Infections of prosthetic joints and related problems. In: Armstrong D, Cohen J (eds). Infectious diseases, pp. 2.44.1–2.44.6. Mosby, London.

Berendt T, Byren I (2004). Bone and joint infection. Clin Med, 4, 510–8. 

Christodoulou C, Gordon P, Coakley G (2006). Polyarticular septic arthritis. BMJ, 333, 1107–8.

Coakley G, et al. on behalf of the British Society for Rheumatology Standards GaAWG (2006). BSR & BHPR, BOA, RCGP and BSAC guidelines for management of the hot swollen joint in adults. Rheumatology, 45, 1039–41.

Courtney P, Doherty M (2009). Joint aspiration and injection and synovial fluid analysis. Best Pract Res Clin Rheumatol, 23, 161–92.

Girdlestone GR (1943). Acute pyogenic arthritis of the hip: an operation giving free access and effective drainage. Lancet 1, 419. [Reprinted in Clin Orthop, 170, 3–7 (1982)]

Kaandorp C. et al. (1997). The outcome of bacterial arthritis: a prospective community-based study. Arthritis Rheum, 40, 884–92. 

Kang SN, Sanghera T, Mangwani J, et al. (2009). The management of septic arthritis in children: systematic review of the English language literature. J Bone Joint Surg Br, 91, 1127–33.

Kaplan SL (2009). Challenges in the evaluation and management of bone and joint infections and the role of new antibiotics for gram positive infections. Adv Exp Med Biol, 634, 111–20. 

Lavy CB, Thyoka M, Pitani AD (2005). Clinical features and microbiology in 204 cases of septic arthritis in Malawian children. J Bone Joint Surg Br, 87, 1545–8.

Lowy FD (1998). Staphylococcus aureus infections. New Engl J Med, 339, 520–9.

Mathews CJ, et al. (2007). Management of septic arthritis: a systematic review. Ann Rheum Dis, 66, 440–5.

Stengel D, et al. (2001). Systematic review and meta-analysis of antibiotic therapy for bone and joint infections. Lancet Infect Dis, 1, 175–88.

Syrogiannopoulos GA, Nelson JD (1988). Duration of antimicrobial therapy for acute suppurative osteoarticular infections. Lancet, ii, 37–40.