Septic arthritis is a term often used to describe patients with an acute pyogenic joint infection due to a bacterial pathogen. The term may also be used in reference to patients with more indolent presentations of infectious arthritis, typically caused by fungi, mycobacteria, or other slow-growing bacteria. In patients who present with an acutely painful and swollen joint, the reported prevalence of septic arthritis is 8% to 27%. Patients with pre-existing inflammatory joint diseases such as gout or rheumatoid arthritis (RA) are at increased risk of septic arthritis, and disease-modifying agents such as corticosteroids and tumor necrosis factor (TNF) inhibitors can alter inflammatory responses, making common bacterial infections more difficult to recognize and infections due to unusual pathogens more common. Patients with septic arthritis need prompt therapy to avoid chronic sequelae and risk of death (often justifying immediate empiric institution of anti-infectives), but there is a general lack of data to support clinical decision making. Thus, patients with septic arthritis provide significant diagnostic and therapeutic challenges, not only for general practitioners, but also for experienced rheumatologists. A proposed diagnostic and treatment algorithm is shown below ( Fig. 20-1 ).
A 71-year-old white woman had a history of breast cancer, arterial hypertension, chronic renal disease, and chronic atrial fibrillation treated with warfarin. One week before admission, pain and swelling of the first metatarsophalangeal (MTP) joint developed during the night. The next day, the joint became red, shiny, and very tender, and the patient experienced chills. In the next few days, the pain, swelling, and redness became more and more intense. She could neither sleep nor walk.
She was evaluated by her primary care physician. The results of laboratory tests revealed elevated inflammatory parameters: the C-reactive protein concentration (CRP) was 131 mg/L (<5) and erythrocyte sedimentation rate (ESR) was 36 mm/h (<15) but her complete blood count (CBC) was normal. Naproxen 500 mg bid was prescribed. In the next 2 days, the symptoms did not improve much and she was sent to the emergency department. On arrival, she reported pain in the left foot. On examination, her body temperature was 36.8 ° C, her blood pressure was 140/82 mm Hg, her pulse was 68 beats per minute, and her oxygen saturation was 98% while she was breathing ambient air. Her left foot and especially her first MTP was hot, red, swollen and extremely tender with restriction of movement. Laboratory tests showed a CRP concentration of 198 mg/L, ESR 74 mm/h, procalcitonin (PCT) 2.72 µg/L (<0.5), and normal values on CBC, liver enzymes and urine analysis. Serum urate and creatinine were slightly elevated at 498 µmol/L (<480), and 135 µmol/L (normal values range: 44-97), respectively. Her chest x-ray study was normal, and her foot x-ray study showed only swelling of the soft tissues around the left first MTP, whereas ultrasound revealed arthritis in the same joint with increased power Doppler signal ( Fig. 20-2 ).
Arthrocentesis of the left first MTP was performed; the aspirate was purulent, with 85,000 white blood cells (WBC)/mm 3 (98% leukocytes and 2% lymphocytes). A Gram’s stain revealed gram-positive bacteria in clusters. Staphylococcus aureus grew from the synovial culture. In addition, monosodium urate monohydrate crystals were detected by polarized light microscopy. Blood cultures were also collected but remained sterile. A diagnosis of infectious arthritis was made, and flucloxacillin 2 g every 6 hours was started intravenously ( Fig. 20-3 ). The foot gradually improved ( Fig. 20-4 ). Two weeks after admission, the patient was discharged free of symptoms. Flucloxacillin was replaced by oral clindamycin 450 mg every 6 hours for 4 additional weeks. Subsequent ambulatory follow-up showed complete resolution of the left first MTP joint septic arthritis.
Discussion of Case Study 1
It is clinically difficult to distinguish a swollen joint due to crystal arthritis from one due to septic arthritis. Yet, this distinction is critical because septic arthritis is associated with a much higher rate of irreversible joint destruction and with an estimated case fatality rate of 11%. Septic arthritis typically presents as a hot, swollen, tender joint with a reduced range of movement. The symptoms are usually present for less than 2 weeks, but significant delays may occur, particularly with low-virulence organisms, tuberculosis (TB), and prosthesis infection. The lower extremities and large joints are most often affected, including the knee in more than 50%, along with the hips, ankles, and wrists.
The patient presented initially with a range of symptoms that suggested an acute gout attack: acute severe pain, redness, swelling at the base of the big toe (first MTP joint, known as podagra) and disability. However, the maximum severity of the gout or pseudogout attack is usually reached over several hours, which was not the case in our patient. Furthermore, oral nonsteroidal anti-inflammatory drugs (NSAIDs) prescribed by the primary care physician did not alleviate the patient’s symptoms. Neither the localization of arthritis nor the absence of fever can be used to exclude septic arthritis.
The peripheral WBC count, ESR, and CRP have limited diagnostic power for discriminating septic arthritis from crystal-induced arthritis, but an elevated serum PCT can be a specific marker for bacterial arthritis, as was the case in our patient. PCT is a propeptide of calcitonin that is produced in the C cells of the thyroid and is a more specific marker of bacterial infection than CRP, rising earlier than CRP and correlating more closely with the severity of the disease.
The most important procedure in the evaluation of monoarthritis is arthrocentesis with microscopic analysis and culture of the synovial fluid aspirated from the affected joint. Importantly, it is generally safe for patients receiving oral anticoagulants to undergo arthrocentesis without a need for alteration of the oral anticoagulant regimen.
Thus, at the onset of acute arthritis, the first MTP was extremely tender and red, and gout seemed a likely diagnosis. NSAIDs therapy failure and worsening of the clinical picture raised the suspicion of septic arthritis. The synovial fluid analysis (e.g., Gram’s stain and culture as well as the detection of monosodium urate crystals) confirmed the diagnosis of concomitant septic and crystal-induced arthritis. Once diagnosed, the treatment consisted of intravenous followed by oral antibiotics, with a favorable clinical outcome.
A 68-year-old woman was admitted to the rheumatology department for the evaluation of subacute oligoarthritis. She had a history of pulmonary hemosiderosis 15 years previously that was believed to be part of an undifferentiated systemic connective tissue disease. A diagnosis of microscopic polyangiitis (MPA) had been made 3 years before the current admission, when she presented with weight loss, malaise, fever, polyarthralgias, myalgias, mononeuritis multiplex and segmental necrotizing glomerulonephritis. Antinuclear antibodies were present at a titer of 1:80 in a speckled pattern. Antineutrophil cytoplasmic antibodies (ANCAs) were positive by immunofluorescence, and antimyeloperoxidase antibodies were positive by enzyme-linked immunosorbent assay (ELISA). She was treated with a monthly pulse of 1 g of cyclophosphamide for 6 consecutive months, followed by oral azathioprine and 48 mg/day of 6-methylprednisolone, which was gradually reduced. The patient was monitored regularly and was clinically stable with the exception of generalized arthralgias which were attributed to concomitant osteoarthritis of the knees and proximal and distal interphalangeal joints. Nine months before the current admission, azathioprine was stopped, but she was still taking 4 mg/day of 6-methylprednisolone. Approximately 3 months before the current admission, the patient experienced the sudden onset of pain in both wrists and fever. She was evaluated at another facility, where a respiratory tract infection associated with bacteremia was diagnosed and treated with 2 weeks of intravenous amoxicillin/clavulanic acid.
Over the course of 2 months, the pain in the right wrist resolved, but the pain and swelling in the left wrist progressed, together with pain and swelling of the right ankle and second proximal PIP of the left hand, prompting referral to our facility.
On examination, her temperature was 37.8 ° C; the head, eyes, ears, neck, chest, lungs and heart revealed no abnormalities. The abdomen was soft, and nontender. The liver and spleen were not palpable. The spine appeared normal for her age. There was a marked atrophy of the interosseal muscles of the hands and tenderness, swelling and warmth over the left wrist ( Fig. 20-5 ), right ankle, and second proximal PIP of the left hand. Osteoarthritic changes were present in other small peripheral joints. The ESR was 67 mm/h, and the CRP was 96 mg/L. The results of a complete blood count were normal, as were levels of blood glucose, electrolytes, urea nitrogen, creatinine, urine and the results of tests for ANCAs, rheumatoid factor, anti-citrullinated peptide antibodies, Lyme antibodies, Chlamydia antibodies and antibodies to streptococcal antigens. Blood cultures were negative. A throat culture for group A beta-hemolytic streptococci was also negative. A plain radiograph of the left wrist showed subchondral bone destruction with subluxations and erosions in the radiocarpal joint and “pencil-in-cup” deformity of the second PIP joint of the left hand. An echocardiogram did not show signs of endocarditis. Arthrocenteses were performed under ultrasound guidance. The synovial fluid from the wrist and ankle was opaque and yellow, with a white-cell count of 95,000 cells/mm 3 (92% polymorphonuclear leukocytes) and negative for the presence of crystals. Synovial fluid from the PIP joint was also purulent. However, Gram’s stains did not show any bacteria.
An incision and drainage of the wrist and repeated needle aspiration of the ankle were performed. Empirical treatment was initiated with ceftriaxone 2 g intravenous (IV) once daily and vancomycin 1 g IV twice daily. Streptococcus pneumoniae, intermediate in susceptibility to penicillin but susceptible to ceftriaxone, was isolated from the synovial fluid culture of all three affected joints. Vancomycin was withdrawn and ceftriaxone was continued for 6 weeks, with marked clinical improvement (with the exception of the wrist, which healed with a poor functional outcome) and a progressive decrease in the ESR and CRP to 21 mm/h and 8 mg/L respectively.
Discussion of Case Study 2
There are several important aspects to this case. The patient, who had a history of systemic vasculitis and had been treated with immunomodulatory drugs, presented with an oligoarthritis that had evolved insidiously for weeks or even months. The initial differential diagnosis of this case was broad and included spondyloarthropathies, reactive arthritis, rheumatic fever, gout, pseudogout, sarcoidosis, Lyme disease, adult-onset Still’s disease, a flare of ANCA-associated vasculitis, viral arthritis, and other diagnoses. However, appropriate examinations, including arthrocentesis and laboratory tests, ruled out many of these possibilities, and a diagnosis of oligoarticular septic arthritis was made.
The clinical and laboratory features of this immunocompromised case, including the oligoarticular small joint involvement, represented a less common presentation of septic arthritis. However, in up to 16.6% of cases of septic arthritis overall, and in 36% of pneumococcal septic arthritis more than one joint is affected. Therefore, an oligoarticular or polyarticular presentation does not exclude septic arthritis. Arthrocentesis of the left wrist, yielding purulent material, and subsequent puncture of other painful and swollen joints led to the correct diagnosis. The evolution of the disease was insidious, probably because of prior treatment with immunosuppressant drugs and amoxicillin. At presentation, she already had radiographic changes showing bone destruction, which is consistent with late septic arthritis (at least 8–10 days after infection).
A 56-year-old female teacher with RA for more than 10 years was treated with methotrexate 20 mg once a week, prednisone 10 mg/day, and hydroxychloroquine 200 mg bid. She had had a right total knee replacement and a right total hip replacement 4 and 2 years previously with good results. An earlier arthroscopy of the left shoulder had been complicated by S. aureus infection, and she had been treated with antibiotics for 4 months. After completing a controlled trial and continuing on the extension trial using a TNF antagonist, she presented to the emergency room with increasing left knee pain and symptoms she believed were related to a flare-up of her RA. Her MCPs were tender, and her left knee showed a small effusion but with an adequate range of motion. Her temperature was 37.2° C. Her laboratory tests included a normal CBC, an ESR of 35 mm/h, and a CRP of 1.9 mg/L. She was given a diagnosis of RA flare, and treated with methylprednisolone 40 mg IM and sent home. Thirty-six hours later, she developed temperature of 38.2° C and pain in multiple joints. She was admitted to the hospital with hypoxia, pulmonary infiltrates, and temperature of 38.9° C. Initially she was treated with IV vancomycin for possible methicillin-resistant S. aureus (MRSA) infection. Cultures of blood and synovial fluids from the left and right knees yielded methicillin-sensitive S. aureus, and vancomycin was changed to nafcillin 12 g/day by continuous infusion. The portal of entry was believed to be multiple excoriations from neurodermatitis. She died on the 8th day.
This case was adapted from the Biologics License Applications (BLA) submitted to the US Food and Drug Administration (FDA) for initial approval of a TNF antagonist.
Discussion of Case Study 3
This case illustrates several clinical situations that are associated with septic arthritis. Septic arthritis should be the first consideration in the differential diagnosis for any patient with monoarticular arthritis, including monoarticular arthritis in the setting of polyarticular disease, or any arthritis accompanied by fever or signs of systemic infection. When a patient with chronic joint disease presents with one or two inflamed joints out of proportion to the rest, this should be an important diagnostic clue to the possibility of septic arthritis. Less typical presentations of septic arthritis may occur when the hosts are at the extremes of the age spectrum; have comorbid disease, prior infections, or prosthetic joints; or are taking therapies that alter the inflammatory and immune responses of the host. In particular, medications that inhibit circulating TNF can contribute to an increased risk for dissemination of infection from a focal source such as a septic joint. Overt signs of inflammation and systemic illness may be muted or absent; history and physical examination of a patient with a swollen joint may not exhibit the presence of an infection. Initial improvement of symptoms of an inflammatory arthritis with use of anti-inflammatory medications does not rule out an underlying infection. A high index of suspicion for infectious arthritis is the key to avoiding morbidity and mortality that might result from diagnostic delay and inappropriate therapies.
ORGANISMS RESPONSIBLE FOR SEPTIC ARTHRITIS
Although almost any pathogen may cause septic arthritis, there are often clinical clues suggesting which might be involved ( Table 20-1 ). S. aureus is the most common pathogen, and is seen frequently in patients with rheumatoid arthritis or other pre-existing joint disease. Because staphylococci often colonize human skin, penetration of the skin (e.g., surgery, trauma, injection drug use) or breakdown of normal skin integrity may result in bacteremic seeding of joints. MRSA septic arthritis most often occurs among persons with health care exposures such as surgery, hemodialysis, hospitalization, or residence in long term care facilities. However, in the past decade, new strains of MRSA have emerged and have caused community acquired infections in a variety of hosts without traditional MRSA risk factors, including healthy children, professional athletes, prison inmates, and injection drug users. Coagulase-negative staphylococci, such as S. epidermidis rarely cause native joint septic arthritis, but are an important cause of infections involving joint prostheses. A variety of streptococcal species commonly cause septic arthritis. S. pyogenes , a common cause of skin and soft tissue infection, is the most frequently isolated streptococcal species causing septic arthritis. Other beta-hemolytic streptococcal species, such as S. agalactiae are often associated with an underlying medical condition such as diabetes, alcoholism, immunosuppression, or malignancy. S. pneumoniae was found to cause 6% of cases of septic arthritis in one review, and was often associated with co-morbid medical conditions, most commonly rheumatoid arthritis.
|Organism||Estimated % of Adult Infections||Clinical Clues|
|Staphylococcus aureus including methicillin-resistant S. aureus (MRSA)||40–70||Healthy adults, skin breakdown, previously damaged joint (e.g., rheumatoid arthritis), prosthetic joint|
|Staphylococcus epidermidis||10–15||MRSA increased in injection drug users and hospital or nursing home-acquired infections|
|Most common agent in early postoperative prosthetic joint infection|
|Streptococcal species||15–20||Healthy adults, splenic dysfunction, usually preceded by primary skin or soft tissue infection|
|Streptococcus pneumoniae||2||Underlying joint disease, prosthetic joint, alcoholism, multiple myeloma, sickle cell anemia, HIV|
|Gram-negative bacteria||5–25||Immune compromised hosts, urinary tract and gastrointestinal infection, children|
|Anaerobic bacteria||1–5||Penetrating trauma|
|Immune compromised hosts|
|Suspect if synovial fluid is foul smelling or air is present in the joint space radiologically|
|Culture should be collected under anaerobic conditions|
|Mycoplasma||Rare||Immune compromised hosts, particularly agammaglobulinemia|
|Dimorphic fungi (Coccidioides, Sporothrix, Blastomyces, Histoplasma,Penicillium)||Rare||Immune compromised hosts|
|Candida species||Rare||Venous catheters, steroids, broad spectrum antibiotics, diabetes, hyperalimentation, trauma|
|Aspergillus species||Rare||Neutropenia, chronic granulomatous disease, organ or bone marrow transplantation|
|Cryptococcus||Rare||HIV, steroids, hematologic malignancy; adjacent cryptococcal osteomyelitis|
|Mycobacterium tuberculosis||Rare||Immune compromised hosts|
|Recent travel to or residence in an endemic area|
|Patients treated with TNF-antagonists|
|Nontuberculous mycobacteria||Rare||Immune compromised hosts|
Infection due to Neisseria gonorrhoeae should be suspected when septic arthritis occurs in a patient with risk factors for sexually transmitted disease. Other gram-negative organisms infect joints infrequently, mostly in the elderly or immunosuppressed, after trauma, or in patients who are injection drug users. Injection drug users have a predilection to develop bacterial arthritis in axial joints, such as the sternoclavicular or sternomanubrial joint. A chronic or indolent presentation should suggest the possibility of infection due to fungi or mycobacteria . Travel history may provide clues to the possibility of infections that are endemic in certain areas, such as coccidioidomycosis in central California and southwestern United States, parts of Mexico, Central America and South America, brucellosis in the Middle East or Burkholderia pseudomallei infection ( melioidosis ) in southeast Asia and northern Australia. A history of penetrating trauma should suggest the potential for polymicrobial infection or infection with gram-negative organisms , anaerobes, and fungi. Pantoea agglomerans is the most common cause of bacterial arthritis caused by penetrating injury due to plant thorns.
The symptoms of joint infections caused by Mycobacterium tuberulosis , or TB of the joint, are nonspecific, and the clinical course is often indolent, usually leading to significant delays in diagnosis, and resultant joint destruction. Pain or local swelling is the most frequent presenting complaint. Fever and weight loss are present in only about 30% of patients with TB arthritis, and only about 50% of patients with joint TB have chest radiographs suggestive of TB infection. Patients who present late in the course of their infection may have evidence of joint destruction including local deformity and restricted range of motion. Some patients with far advanced disease have draining sinuses. The most significant step toward diagnosing TB joint infections is to consider the possibility of the diagnosis in the appropriate clinical setting. Often, a tuberculin skin test is performed. However, this test is of limited use in determining active disease and is best used for screening for latent infection in high-risk populations.
The rheumatologist should anticipate increases of TB arthritis with the global increase of TB related to continued immigration and travel and the increased incidence and dissemination of TB and fungal infections attendant to biologic therapies that antagonize TNF. Bone and joint infection may account for 10% to 35% of cases of extrapulmonary TB and for almost 2% of all cases of TB. TB cases reported in association with TNF antagonists show dissemination of TB to extrapulmonic sites in 35% to 50%. Tuberculous skeletal infection can develop either by hematogenous or lymphatic spread from chronic pulmonary or lymph node foci, or by reactivation of latent mycobacteria at the site seeded in the primary infection. When local immune defenses fail, as with poor nutrition, advancing age, human immunodeficiency virus infection, or renal impairment, reactivation with a progression to clinically apparent disease may occur. Prolonged corticosteroid use may play an important role in the development of TB arthritis. In a recent study from Taiwan, 16 of 51 patients (31.4%) had long-term corticosteroid use before developing tuberculous arthritis. Vigilance regarding TB infection has also been recommended during treatment with leflunomide. A large US-based study of all cases of bone and joint TB over a 4-year period revealed that the most common site of bony TB was the spine, followed by weight-bearing joints (hip and knee) and other sites.
The spectrum of Mycobacterium leprae infections reflects the diversity of the host cellular immune responses. Rarely, foamy macrophages (lepra cells) can be seen in the synovial fluid and the tissue can show a spectrum of lymphocytic infiltrate and vasculitis with the lepra reaction or pannus with macrophage granulomas.
Atypical or nontuberculous mycobacteria are normal commensals in water, cattle milk, soil, and dust, and rarely may infect the joint of someone chronically exposed or whose immune system has been compromised by disease or therapy. Infections with Mycobacterium avium complex, Mycobacterium kansasii, or Mycobacterium marinum have been reported in dairy farmers, pool cleaners, patients with scleroderma and polymyositis, and in prosthetic joints. The diagnosis has usually required great diligence and multiple biopsies, which are fastidiously cultured and occasionally complemented by polymerase chain reaction (PCR) studies. Most of these organisms are sensitive to ethambutol, rifampin, and other combinations.
Disseminated coccidioidomycosis can affect the musculoskeletal system, especially in RA patients taking TNF antagonists who frequent endemic areas. The organisms can be found in multiple, noncontiguous sites of the axial skeleton, the paraspinal muscles, the appendicular long bones, and the synovial joints of the lower extremities, especially the knee.
Other fungal joint infections include blastomycosis , a rare systemic fungal disease endemic to the Great Lakes, Mississippi and Ohio River Valleys, northwest Ontario and Manitoba involving the skin, lung, bones and joints, and disseminated Cryptococcosis, reported in the immunosuppressed, including those treated with TNF inhibitors.
HOST FACTORS IMPORTANT IN SEPTIC ARTHRITIS
As illustrated in Case Studies 2 and 3, patients with systemic rheumatic diseases treated with immunosuppressive therapies carry an elevated risk of septic arthritis. Of these diseases, RA is the one most commonly associated with septic arthritis, probably because of the combination of joint damage and immunosuppressive therapy. Risk factors for septic arthritis in patients with RA are advanced age (≥ 80 years); history of chronic diseases, including diabetes; superficial skin ulceration or other minor primary infection, oral carriage of S. aureus , the presence of a prosthetic joint, and biologic agent therapy.
The infections encountered in patients on TNF-blocking therapies range from localized to disseminated. It appears that not only is the incidence of certain infections increased with TNF blockade therapy, but the ability to contain these infections is also impaired .
TNF inhibitors increase the susceptibility to TB and reactivation of latent TB infections. The majority of patients who take these drugs have RA, and although RA itself has been associated with the development of TB, TNF inhibitors clearly increase the risk of TB beyond that which may be associated with RA itself. Recent studies have suggested a greater risk of TB for anti-TNF monoclonal antibodies than for the soluble TNF receptor construct, particularly with regard to reactivation of latent infection. TB occurring in association with TNF inhibitors has a higher likelihood of involving extrapulmonary sites and of being disseminated at presentation, and this includes TB involving the bone or joint.
Although pulmonary and extrapulmonary TB have received the most attention, a wide range of infectious diseases have been observed in the setting of TNF blockade treatment. The spectrum of pathogens causing invasive disease in patients receiving TNF blockade therapy ranges from common gram-positive and gram-negative bacteria to more opportunistic organisms such as M. tuberculosis, Cryptococcus, and Aspergillus.
An explanation for the increased risk of infections with common organisms such as S. aureus in patients treated with anti-TNF could be the demonstration that S. aureus and other gram-positive bacteria are potent inducers of TNF secretion from macrophages and that a local increase of TNF levels might improve host defenses against staphylococcal foreign body infections. To limit anti-TNF–associated infections, antimicrobial prophylaxis has been discussed by some authors because it has been shown to be effective in the transplantation arena. However, the last updated consensus statement on biologic agents for the treatment of rheumatic disease does not support this concept. Patients on anti-TNF therapy should be instructed to seek medical attention in case of unexplained fever, shaking chills, or sudden arthritis of one or more joints in otherwise quiescent disease.
DIAGNOSTIC WORK-UP OF THE PATIENT WITH SUSPECTED INFECTIOUS ARTHRITIS
Joint Fluid Analysis
Prompt and specific diagnosis of septic arthritis demands the recovery of synovial fluid by joint aspiration. For cases in which joint fluid analysis is either not possible or is nondiagnostic, synovial biopsy for culture and special pathology studies may be required. Ultrasound, fluoroscopy, or computed tomography (CT) may be needed to guide a successful arthrocentesis in cases in which joint fluid is scant or loculated, or the affected joint is difficult to access, such as the hip or a spinal facet joint. Synovial fluid Gram’s staining and microscopy give a positive result in 50% of cases of septic arthritis, and subsequent fluid culture increases the yield, but a joint can be septic even in the absence of positive microscopy or culture.
The synovial fluid specimen should be obtained using sterile technique, ideally before starting antibiotic therapy, and should be sent immediately to the laboratory for cell count and differential, Gram’s stain, and culture. It is also important to examine synovial fluid for the presence of crystals, but their presence does not exclude the possibility that an infectious arthritis is also present.
The synovial fluid WBC in septic arthritis, crystal arthritis, and other causes of joint inflammation overlap considerably. However, a higher synovial fluid WBC and a greater proportion of polymorphonuclear leukocytes increase the likelihood of a diagnosis of septic arthritis. WBC counts of less than 25,000 cells/mm 3 , more than 25,000 cells/mm 3 , more than 50,000 cells/mm 3 and more than 100,000 cells/mm 3 give a septic arthritis likelihood ratio of 0.32, 2.9, 7.7 and 28, respectively. On the other hand, in a recent study, it was shown that a synovial fluid WBC count cut off at 50,000 cells/mm 3 lacks the sensitivity required to be clinically useful in ruling out infective arthritis. Likewise, in crystal-induced arthritis a synovial fluid WBC count of 30,000 to 50,000 cells/mm 3 is typical, but even counts of 100,000 cells/mm 3 are sometimes observed. In the case of a prosthetic joint synovial fluid, a WBC count of more than 17,000 cells/mm 3 and a differential of more than 65% polymorphonuclear leukocytes had a sensitivity for diagnosing infection of 94% and 97% and a specificity of 88% and 98%, respectively.
The synovial fluid glucose is often depressed and the lactic acid concentration is elevated in the synovial fluid with infection. Because lactic acid levels also rise in joint inflammation through any cause, these tests are not sufficiently sensitive to be of widespread diagnostic use.
Gram’s stain of the synovial fluid may help narrow or expand the differential diagnosis, enable a rapid presumptive diagnosis, and influence early antibiotic therapies before culture results and sensitivities return. For example, broad-spectrum empiric therapy targeted at gram-negative pathogens would not generally be necessary if the Gram’s stain shows gram-positive cocci in clusters indicative of a staphylococcal infection. Or if the clinical history suggests a urinary tract source and the Gram’s stain shows gram-negative rods, empiric antistaphylococcal therapy could be avoided and therapy targeted at urinary tract organisms started early. Although the Gram’s stain is more often positive in infections due to gram-positive organisms (71%), the overall sensitivity of a Gram’s stain for diagnosis of septic arthritis is only 29% to 50%. Therefore, a negative Gram’s stain should not be taken as evidence against an infectious cause of arthritis.
In patients with suspected arthritis due to M. tuberculosis, stains of joint fluid for acid-fast bacilli (AFB) (e.g., Gram’s and fluorescent auramine stains) are positive in a minority of cases, although a higher frequency of smear-positive disease has been reported in Taiwan.
Some fastidious organisms may fail to grow if transit time to the laboratory is prolonged or the specimen is otherwise mishandled. Even hearty organisms such as S. aureus may fail to demonstrate significant growth on conventional media if the number of organisms in the sample is small or antibiotics have been given. Several studies have demonstrated that there is an increased yield of cultured organisms when joint fluid is inoculated into various types of blood culture bottles. The recognition of Kingella kingae as an important cause of septic arthritis in children has been attributed to this practice. Although its cost-effectiveness has not been adequately studied, routine use of blood culture bottles for culturing joint fluid should be considered. Often, however, joint fluid is sent directly to the microbiology laboratory for direct plating on conventional agar. In at least one study, the choice of culture method was less critical if the sample was obtained and preserved with caution, with attention to aseptic techniques throughout and immediate transport to the laboratory. If this is done, the largest volume of specimen obtainable should be sent in a recapped syringe after removing the needle and excess air. If it is necessary to separate the specimen for transport to different laboratories, a green-top tube containing sodium heparin can be used for the portion sent for routine microbiology, as long as the volume of joint fluid is sufficient (1–2 mL); high concentrations of heparin may inhibit growth of some bacteria.
Although many organisms that cause septic arthritis grow on conventional media, other significant pathogens are difficult to culture and may require special culture techniques. TB and other mycobacterial infections, fungal infections, and anaerobic infections are difficult to detect from synovial fluid analysis, and diagnosis may depend on a synovial biopsy or PCR. For this reason, it is important that the clinician communicate effectively with the microbiology laboratory concerning the most likely pathogens suspected based on the clinical scenario. For example, alerting the laboratory to the search for Neisseria will have the fluid plated on lysed blood (chocolate) agar. To isolate Borrelia burgdorferi (Lyme disease) synovial fluid has to be cultured in Barbour-Stroenner-Kelley media (BSK) (Sigma B3528). A history of chronic arthritis and pneumonia might lead a suspicion of mycobacterial or fungal arthritis and prioritization of plating the specimen on growth media for mycobacteria or fungi when specimen volume is limited.
Early consultation with an infectious disease specialist can help broaden the microbiologic differential diagnosis and provide a liaison with the microbiology laboratory to achieve efficient, yet thorough testing.
For the diagnosis of mycobacterial joint infections it is important to culture material from deep structures, such as bone, abscesses, synovial fluid, or synovial tissue rather than culturing drainage fluid, because these specimens can grow colonizing organisms, such as bacteria or fungi, which may cloud the diagnosis. An older review of the use of synovial fluid culture for M. tuberculosis reported a sensitivity of 79%, whereas synovial tissue culture had a sensitivity of 94%. However, to culture tubercle bacilli on the Lowenstein-Jensen medium requires 4 to 8 weeks to detect growth. Substantial improvement in the time taken for detection and total number of positive cultures can be gained by using a carbon-14 broth based growth system. Furthermore, histology can be very useful in suggesting the diagnosis, and suspected TB is one of the few indications for synovial biopsy.
Quantitative molecular methods such as the PCR are still emerging but may in the future facilitate the rapid detection of pathogenic organisms. As an example, a novel adaptation of a probe-based real-time PCR assay targeting the 16S rRNA gene for early and accurate diagnosis of bacterial arthritis has been described. The assay algorithm consisted of initial broad-range eubacterial detection, followed by Gram typing and species characterization of the pathogen. One hundred and twenty-one clinical synovial fluid samples from patients presenting with suspected acute septic arthritis were tested. The sensitivity and specificity of the assay were 95% and 97%, respectively, versus synovial fluid culture results. Gram-typing probes correctly identified 100% of eubacterial positive samples as to gram-positive or gram-negative status, and pathogen-specific probes correctly identified the etiologic agent in 16/20 eubacterial positive samples. The total assay time from sample collection to result was 3 hours. Further large-scale prospective studies are required for clinical validation of this approach.
The Importance of Blood Cultures
Most cases of septic arthritis are the result of hematogenous spread of microorganisms to the synovial membrane of one or more joints. In one study, specific culture diagnosis was made by blood culture alone in 12.5% of cases. Therefore, blood cultures are an essential part of the initial diagnostic evaluation in patients with suspected septic arthritis, and whenever possible, the clinician should obtain at least two sets of blood cultures before initiating antibiotic therapy. In general 27% to 33% of patients with septic arthritis have a positive blood culture, and even more frequently in specific infections such as those caused by Streptococcus pneumoniae (77%). In the absence of a history of penetrating trauma, surgery or pre-existing contiguous osteomyelitis, a hematogenous source of the infection should be assumed, and a source aggressively sought. The history and physical examination should identify potential portals of entry of microorganisms into the bloodstream, such as invasive devices or surgeries, skin lesions (e.g., dermatitis, ulcerations, lacerations, excoriations), mucosal or gastrointestinal tract disruption (e.g., dental abscesses, pneumonia, inflammatory bowel disease, gastroenteritis, malignancy), and other risk factors for primary bacteremia, such as alcoholism, cirrhosis, multiple myeloma, or injection drug use. If an overt bloodstream infection is detected, repeat blood cultures should be obtained after 48 hours of specific therapy to ensure that the blood has been sterilized. Persistent bacteremia or fungemia suggests that the patient has an endovascular infection (endocarditits, septic thrombophlebitis, mycotic aneurysm, infected venous catheter) or other persistent un-drained focus of infection. Consultation with an infectious disease specialist should be considered, because this has been shown in six different studies to reduce morbidity, mortality, and cost of care in patients with bacteremia.
There are no robust studies prospectively investigating the optimal imaging methods to be used in the diagnosis of septic arthritis. Plain radiography, although not useful for imaging the joint in an acute onset of arthritis, should be performed routinely to exclude an underlying trauma, tumor, or osteomyelitis. Ultrasound (US) is used in daily clinical practice to assess joints with inflammatory arthritis and can clearly demonstrate a fluid collection in the joint or in the periarticular tissue (e.g., bursae). Moreover, US-guided needle placement in the joint, for example, the shoulder or hip, is more accurate and less painful than needle placement without US guidance. Scintigraphy cannot distinguish septic arthritis from other causes of joint arthritis. A study on MRI findings described that the enhancement of proliferating synovia, perisynovial edema, and joint effusions would correlate most accurately with the presence of an infection, however, no sterile inflammatory controls were included in the study design. Thus, no radiologic technique is currently sufficiently specific to be diagnostically useful in suspected septic arthritis. On the other hand, the ability of MRI to define an osteomyelitis, pyomyositis, or cellulitis may permit optimal treatment, particularly if performed before any interventional procedures.