S e p tic A r t h r i t i s o f N ati v e

Joints
John J. Ross, MD

KEYWORDS
 Septic arthritis  Arthrocentesis  Synovial fluid  MRSA

KEY POINTS
 Septic arthritis is a true rheumatologic emergency that may lead to disability or death.
Prompt evacuation of the joint, either by arthrocentesis at the bedside, open or arthro-
scopic drainage in the operating room, or imaging-guided drainage in the radiology suite,
is mandatory.
 Arthrocentesis is diagnostic and therapeutic in patients with suspected septic arthritis.
Patients should be treated empirically for septic arthritis if the synovial fluid white blood
cell (WBC) count exceeds 50,000 cells/mm3. Patients with bacterial arthritis who are debil-
itated or immunosuppressed may have lower synovial fluid WBC counts.
 Methicillin-resistant Staphylococcus aureus (MRSA) is a major cause of septic arthritis in
the United States; however, it is less prominent as a cause of septic arthritis in Europe.
MRSA joint infection seems to be associated with worse outcomes. This may be due to
its greater virulence, delays in initiation of appropriate antibiotics, and the older age of
many affected patients.
 Antibiotic courses of 3 to 4 weeks in duration are generally adequate for uncomplicated
bacterial arthritis. Treatment duration should be extended to 6 weeks if there is imaging
evidence of accompanying osteomyelitis.

PATHOGENESIS

Septic arthritis of native (nonprosthetic) joints is uncommon but not rare, with approx-
imately 2 cases per 100,000 people per year.1 Occult bacteremia is probably the usual
cause. Synovium is a vascular tissue that lacks a protective basement membrane,
making it vulnerable to bacteremic seeding.2 Minute breaks in the skin or mucous
membranes may allow staphylococci and streptococci to gain access to the blood-
stream. Gram-negative septic arthritis may arise from bacteremia from injection
drug use or loss of integrity of the gastrointestinal or urinary tracts. Occasionally, sep-
tic arthritis is the direct result of penetrating trauma, such as human or animal bites or

The author has no disclosures.

Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, 15 Francis
Street, PBB-B420, Boston, MA 02115, USA
E-mail address: [email protected]

Infect Dis Clin N Am 31 (2017) 203–218

http://dx.doi.org/10.1016/j.idc.2017.01.001 id.theclinics.com
0891-5520/17/ª 2017 Elsevier Inc. All rights reserved.
204 Ross

errant injection drug use. This is the most common means of infection of the small
joints of the hands and feet.3 Rarely, bacterial arthritis may arise as a complication
of arthroscopy or therapeutic joint injection with corticosteroids.
Most cases of septic arthritis are caused by gram-positive organisms. Enteric gram-
negative rods account for 43% of community-acquired bacteremias, but cause only
10% of septic arthritis.4,5 This likely relates to the superior ability of gram-positive organ-
isms to bind connective tissue and extracellular matrix proteins. Staphylococcus aureus,
the commonest cause of septic arthritis, produces several surface adhesins that bind to
extracellular matrix proteins. Staphylococcal strains defective in microbial surface com-
ponents recognizing adhesive matrix molecules are less arthritogenic in animal models.6
Joint damage in septic arthritis results from bacterial invasion, host inflammation,
and tissue ischemia. Bacterial enzymes and toxins are directly injurious to cartilage.
Cartilage may suffer ‘‘innocent bystander’’ damage, as host neutrophils release reac-
tive oxygen species and lysosomal proteases. Cytokines activate host matrix metal-
loproteinases, leading to autodigestion of cartilage.7 Ischemic injury also plays a
role. Cartilage is avascular, and highly dependent on diffusion of oxygen and nutrients
from the synovium. As purulent exudate accumulates, joint pressure increases, and
synovial blood flow is tamponaded, resulting in cartilage anoxia.8

RISK FACTORS

The most robust risk factor for septic arthritis is preexisting joint disease (Box 1). Up to
47% of patients with bacterial arthritis have prior joint problems.3 A high index of sus-
picion for septic arthritis should be maintained in patients with other rheumatologic
conditions, such as rheumatoid arthritis (RA), osteoarthritis, gout, pseudogout, recent
trauma, prior joint surgery, and systemic lupus erythematosus. Of these, RA is the
most common, and is associated with worse outcomes.
Patients with RA are at high risk for bacterial arthritis from the combination of joint
damage, immunosuppressive medications, and poor skin condition. Polyarticular

Box 1
Risk factors for septic arthritis of native joints

Preexisting joint diseases

Rheumatoid arthritis
Gout and pseudogout
Osteoarthritis
Lupus
Trauma
Recent surgery
Diabetes mellitus
Intravenous drug use
Cirrhosis
End-stage renal disease
Prednisone and other immunosuppressive medications
Skin diseases
Psoriasis
Eczema
Skin ulcers
Human bite (fight bite)
 Septic Arthritis of Native Joints 205

disease and complications are common, functional outcomes are worse, and mortality
is high in patients with RA with septic joints.9 Diagnosis is often delayed because of the
confusion of septic arthritis for a flare of RA.
In gouty patients with septic arthritis, inflammation results in shedding of synovial
microtophi, obscuring the diagnosis. All patients with apparent gouty joints should
also have routine Gram stain and culture performed to exclude concomitant septic
arthritis.10,11
Two other broad categories of risk for septic arthritis are conditions causing loss of
skin integrity, such as psoriasis, eczema, skin ulcers, and injection drug abuse, and
conditions associated with compromised immunity, such as diabetes mellitus, renal
failure, cirrhosis, and immunosuppressive drugs. Tumor necrosis factor blockers dou-
ble the risk of septic arthritis in RA, and are associated with septic arthritis with intra-
cellular pathogens, such as Listeria and Salmonella.12 However, up to 22% of patients
with septic arthritis have no medical risk factors and no underlying joint disease.13
A review of cases of septic arthritis from the National Hospital Discharge Survey
revealed that the average age of patients with septic arthritis had risen from 37 years
in 1979 to 51 in 2002. In recent years, patients had more comorbid medical conditions,
had septic arthritis in the setting of complicated hospitalizations, and were more likely
to have infection with antibiotic-resistant organisms. Overall in-hospital mortality was
2.6%, and was similar from 1979 to 2002.14

CLINICAL PRESENTATION

The diagnosis of septic arthritis is straightforward in the classic patient with fever,
rigors, and a warm, swollen, and exquisitely painful joint. Unhappily, however, the clin-
ical and laboratory diagnosis of septic arthritis is often imprecise. High-grade fever is
present in only 58%,13 although 90% have at least low-grade fever.15 Serum leukocy-
tosis is present in only 50% to 60% of patients.13,15 Joint pain is blunted in the immu-
nosuppressed, such as the RA patient on corticosteroids, leading to delayed
diagnosis and more complications.16
Predictors of mortality in a multivariate analysis include age older than 65 years,
confusion at presentation, and polyarticular disease. Predictors of joint damage
include age older than 65 years, diabetes mellitus, and infection with b-hemolytic
streptococci.13

JOINT DISTRIBUTION

The knee is the principal target of bacterial septic arthritis. Forty-five percent of septic
arthritis cases in adults involve the knee.3 Presumably, this is due to the imperfect hu-
man adaptation to bipedal locomotion. The enormous mechanical stresses about the
knee joint particularly predispose it to injury.17 Other large joints of the appendicular
skeleton, including the hip (15%), ankle (9%), elbow (8%), wrist (6%), and shoulder
(5%), are commonly involved in adults.3 Polyarticular disease is seen in approximately
10% to 20% of cases. It is more common with gonococcal, pneumococcal, group B
streptococcal, and gram-negative septic arthritis. Polyarticular septic arthritis is usu-
ally asymmetric, and involves an average of 4 joints. At least 1 knee is involved in 72%
of cases. Major risk factors are steroid therapy, RA, lupus, and diabetes mellitus.18

SEPTIC ARTHRITIS OF CARTILAGINOUS JOINTS

Involvement of cartilaginous joints of the axial skeleton is infrequent, except in intrave-

nous drug users.19 Septic arthritis in intravenous drug users often affects the pubic
206 Ross

symphysis and the sternoclavicular and sacroiliac joints. These joints are uncommonly
infected in other patients with septic arthritis.20–23
The sacroiliac joint has a mobile synovial portion and an immobile fibrocartilaginous
portion (syndesmosis). Sacroiliac septic arthritis is generally seen in younger patients,
although cases in the elderly are sporadically observed. Patients present with buttock
pain and fever. It may be difficult to localize the pain on examination. The FABERE test
(Flexion, ABduction, External Rotation, and Extension) stresses the sacroiliac joint. It is
performed on the supine patient by placing the ipsilateral medial malleolus on the
opposite knee. The ipsilateral knee is depressed, with pressure exerted on the oppo-
site superior iliac spine. Pain is elicited in sacroiliitis, although the FABERE test is not
specific for infection.20,21
The pubic symphysis is a nonsynovial cartilaginous joint. Infection of the pubic sym-
physis presents with fever, suprapubic and hip pain, and a waddling, antalgic gait. Pu-
bic symphysis septic arthritis is rare. Most cases occur in patients with well-defined
risk factors, such as intravenous drug use, pelvic malignancy or surgery, and athletes,
such as soccer players, who are prone to overuse injuries of the hip adductors and pu-
bic periostitis.22
The sternoclavicular joint is a synovial joint, containing a fibrocartilaginous disc that
separates it into 2 compartments. Infection of the sternoclavicular joint may account
for up to 17% of septic arthritis in intravenous drug users. This joint is likely infected
from phlebitis or valvulitis of the underlying subclavian vein, after injection of contam-
inated drugs into the upper extremity. Computed tomography (CT) or MRI should
be obtained routinely in patients with sternoclavicular septic arthritis, given the high
frequency of complications requiring surgery, such as chest wall abscess or
mediastinitis.23

BACTERIOLOGY
Staphylococcus aureus
The most common cause of septic arthritis is Staphylococcus aureus, which accounts
for 52% of cases (Table 1). Only 46% have an underlying focus of staphylococcal
infection, such as cellulitis.24 In the remainder, septic arthritis presumably arises as
a consequence of transient bacteremia from a skin or mucous membrane source. Out-
comes may be poor. Mortality ranges from 7% to 18%, and osteomyelitis or loss of
joint function occurs in up to 27% to 46%.13,15
Methicillin-resistant S aureus (MRSA) is increasing in importance in septic arthritis,
especially in the United States. In a recent series in Boston, 25% of septic arthritis
cases were caused by MRSA.25 All cases were associated with chronic illness, older
age, and health care exposure. A high proportion of cases due to MRSA also has been
seen in series from Detroit (21%) and northern California (50%).26,27 MRSA caused
36% of septic arthritis in a recent series from São Paolo, Brazil, and 42% in a series
from in Taiwan.28,29 MRSA is less common as a cause of septic arthritis in Europe,
causing 15% of cases in a French series, and only 5% of Swiss cases.30,31

b-hemolytic Streptococci
Group B streptococci (Streptococcus agalactiae) have emerged as invasive patho-
gens in the elderly, especially those with diabetes, cirrhosis, and neurologic disease.32
At one center, group B streptococci increased from 1% of septic arthritis cases in the
1980s, to 6% in the 1990s.33 In 2 other recent series, group B streptococci caused
10% of septic arthritis.34,35 Bacteremia was seen in 66%, polyarticular disease in
32%, and mortality was 9%.34 Other b-hemolytic streptococci, particularly group A
 Septic Arthritis of Native Joints 207

Table 1
Microbiology of 505 cases of septic arthritis in large series reporting data from 1999–2013

Bacteria Number (%)

Staphylococci 282 (56)
Methicillin-sensitive Staphylococcus aureus (MSSA) 214 (42)
Methicillin-resistant S aureus (MRSA) 51 (10)
Coagulase-negative staphylococci 17 (3)
Streptococci 83 (16)
Unspecified streptococcal species 56 (11)
Viridans streptococci 7 (1)
Streptococcus pneumoniae 5 (1)
Other streptococcal species 15 (3)
Gram-negative rods 78 (15)
Pseudomonas aeruginosa 30 (6)
Escherichia coli 14 (3)
Proteus species 5 (1)
Klebsiella species 5 (1)
Others 21 (4)
Others 62 (12)
Polymicrobial 25 (5)
Anaerobes 3 (0.6)
Mycobacterium tuberculosis 9 (1.8)
Neisseria gonorrhoeae (gonococcus) 6 (1.2)
Miscellaneous 19 (4)

Data from Refs.26,29–31

streptococci (Streptococcus pyogenes), may cause an equally virulent form of septic

arthritis in adults, although with lesser frequency.

Pneumococcus
Streptococcus pneumoniae caused 6% of cases of septic arthritis in an older literature
review,5 but it appears to be declining in importance as a cause of septic arthritis. In a
more recent series, only 3% of septic arthritis was pneumococcal.36 This may be
attributable to increased herd immunity from childhood pneumococcal vaccination,
which has been associated with fewer cases of invasive pneumococcal disease in
adults.37
As with septic arthritis caused by group B streptococci, pneumococcal septic
arthritis is notable for a high frequency of bacteremia and polyarticular disease.
Only 50% have an underlying focus of pneumococcal disease, such as pneumonia.
Mortality in adults is high (19%), although functional outcomes are good in 95% of sur-
vivors. Drug resistance may be an increasing problem.5

Gonococcal Arthritis
Neisseria gonorrhoeae, once the leading cause of septic arthritis in young adults in the
United States, has dwindled in importance since the 1980s, and now accounts for only
1% of septic arthritis cases. However, recent increases in high-risk sexual behavior
208 Ross

and rising rates of gonorrhea resistant to fluoroquinolones and azithromycin suggest

that epidemic gonorrhea could be poised for a comeback.
Gonococcal septic arthritis is a distinct clinical syndrome, with a good prognosis.
Seventy-five percent of cases occur in women; menses and pregnancy increase the
risk of disseminated gonococcal infection38; 72% of cases are polyarticular.24 Patients
may experience a fleeting and migratory polyarthritis, or a more conventional septic
arthritis picture of several hot, swollen, and exquisitely tender joints. Knee involvement
is most common. The characteristic hemorrhagic pustules of disseminated gono-
coccal infection are found in 42% of patients, and tenosynovitis in 21%. Urinary signs
or symptoms are present in only 32%. Gonococci are recovered from joint fluid in less
than 50% of cases. This is probably largely because of the difficulty in recovering
these fastidious organisms from cultures, but it may also indicate that some cases
of gonococcal arthritis are immune-mediated.38 Complete recovery is the rule with
appropriate therapy, and sequelae are rare.

Enteric Gram-Negative Rods

Gram-negative rods cause approximately 10% to 15% of adult septic arthritis. Two
major groups are at risk: elderly patients with comorbid medical conditions, and young
intravenous drug users.39,40 Outcomes are typically better in the latter group. Recent
data suggest that outcome of gram-negative septic arthritis in older patients may be
relatively good with prompt diagnosis and aggressive therapy, with mortality rates of
only 5%, and poor functional outcomes in 32%. Perhaps surprisingly, an underlying
source of gram-negative septic arthritis, such as a urinary tract infection, is found in
only 50% of older patients.41 Gram-negative bacilli have waned in importance as a
cause of septic arthritis in injection drug users in recent years.

Meningococcal Arthritis
The incidence of arthritis in invasive meningococcal disease is as high as 14%. The
usual presentation is a monoarthritis or oligoarthritis, involving the knee and other large
joints. In most cases, arthritis develops several days into antibiotic therapy, and the joint
fluid is sterile. Synovial fluid immune complexes have been detected in several of these
patients, suggesting an immunologic basis of arthritis.42,43 Less commonly, patients
present with an isolated septic joint (primary meningococcal arthritis), or an arthritis-
dermatitis syndrome akin to disseminated gonococcal infection.44 Joint outcomes in
meningococcal arthritis are usually excellent, with complete preservation of function.42

SEPTIC ARTHRITIS IN SPECIFIC POPULATIONS OR AFTER UNUSUAL EXPOSURES

Coagulase-Negative Staphylococci and Postarthroscopic Septic Arthritis
Most isolates of coagulase-negative staphylococci from native joints are contami-
nants, but they can be true joint pathogens after arthroscopy, anterior cruciate liga-
ment reconstruction, and other orthopedic procedures. Septic arthritis after
arthroscopy is associated with intra-articular corticosteroids, longer operating times,
multiple prior procedures, and chrondroplasty or soft tissue debridement. Coagulase-
negative staphylococcal joint infection tends to present in an indolent fashion, with
low-grade fever, normal peripheral leukocyte counts, and mild to moderate joint
symptomatology. Two weeks of parenteral antibiotics are usually curative for postar-
throscopic septic arthritis.45 Septic arthritis complicating anterior cruciate ligament
reconstruction is treated with aggressive arthroscopic irrigation and debridement, fol-
lowed by antibiotics for 6 weeks. Grafts usually can be retained, if they appear intact
on arthroscopic inspection.46
 Septic Arthritis of Native Joints 209

Mycobacterium Tuberculosis
Tuberculosis causes up to 2% of septic arthritis. A high clinical suspicion is required to
diagnose mycobacterial joint infection. Compared with typical septic arthritis, which
presents with a hot and exquisitely painful joint, tuberculous arthritis usually pursues
an indolent course. Patients often have gradually progressive joint pain and swelling
that mimics osteoarthritis. Symptoms are often present for more than a year before
diagnosis. Fever and weight loss are usually absent, and only 50% of patients have
chest radiographs suggestive of tuberculous infection. Because delayed diagnosis
is typical, abscesses, cutaneous fistulae, osteomyelitis, and joint deformities are often
present. Synovial biopsy for histology and mycobacterial culture is the highest yield
diagnostic procedure. Treatment of tuberculous arthritis is similar to that of tubercu-
losis at other sites. For drug-sensitive disease, regimens lasting 6 to 9 months are usu-
ally curative.47

Brucellosis
Brucella species are a common cause of subacute or chronic arthritis in countries in
which livestock are not vaccinated and unpasteurized dairy products are consumed.
In a recent Israeli series, 11% of septic arthritis was due to brucellosis, with most
cases occurring in rural Arab populations.48 In the United States, most cases are
seen in immigrants from Latin America and the Middle East. The sacroiliac joint is
involved in up to 54% of patients, for unclear reasons. Spondylitis occurs in 7%. Other
patients present with monoarthritis or oligoarthritis, with lower extremity predomi-
nance. It has been suggested that brucellar arthritis may be reactive.49 Brucellosis
is diagnosed by blood culture or serology. The highest cure rates were reported
with the combination of doxycycline for 45 days and streptomycin for 14 days.50

Foreign Body Synovitis and Plant Thorn Injuries

Penetrating joint injuries involving thorny plants, sea urchin spines, wood splinters,
and other foreign bodies may cause a chronic synovitis. This arthritis is usually allergic
in nature. The plant pathogen Pantoea agglomerans, however, has been implicated in
several cases of thorn injuries.51 There is also a report of Nocardia asteroides infection
in this setting.52 Thorn fragments and other synovial foreign bodies are best localized
by ultrasound. Treatment is arthrotomy, with complete resection of foreign matter, and
appropriate antibiotics based on joint cultures.

Human and Animal Bites

Human bites cause polymicrobial infection, involving aerobic bacteria, such as staph-
ylococci; a- and b-hemolytic streptococci; oral gram-negative rods, such as Eikenella
corrodens; and oral anaerobes, including Prevotella, Fusobacterium, and Peptostrep-
tococcus species.53 Animal bites have a similar bacteriology, with Pasteurella multo-
cida as an important additional pathogen.54,55 Treatment relies on drainage and
debridement of devitalized tissues, careful assessment for tenosynovitis, and use of
ampicillin-sulbactam, or a regimen with similar aerobic and anaerobic activity.

Whipple Disease
The multisystem disorder, Whipple disease, is caused by the fastidious actinomycete
Tropheryma whippelii. In 63% of cases, a migratory, nondestructive, peripheral
arthritis is the initial manifestation, preceding the onset of abdominal pain, diarrhea,
malabsorption, and weight loss by a mean of 8 years in 1 series. Because patients
often have HLA-B27 positivity, the arthritis may be mistakenly treated with
210 Ross

immunosuppressive medication, precipitating gastrointestinal symptoms.56,57 Fever,

lymphadenopathy, hyperpigmentation, and cardiac and neurologic involvement also
may be prominent. Diagnosis is usually made on the basis of small bowel biopsy,
although polymerase chain reaction of synovial fluid also may be useful.58 Two weeks
of parenteral ceftriaxone is recommended as initial therapy, followed by oral
trimethoprim-sulfamethoxazole for at least 1 year.59

Mycoplasmas and Ureaplasmas

True septic arthritis caused by mycoplasmas and ureaplasmas is unusual, occurring
almost exclusively in the setting of hypogammaglobulinemia or organ transplantation.
Diagnosis depends on polymerase chain reaction or cultivation on special media.
Good outcomes have been reported with the combination of doxycycline and quino-
lones over weeks to months.60–62

Intravenous Drug Users

Before 1983, Pseudomonas aeruginosa caused 64% of reported septic arthritis in
intravenous drug users, with S aureus responsible for only 10%. After 1983, the roles
of these pathogens were reversed: P aeruginosa resulted in only 9% of septic arthritis
in intravenous drug users, whereas S aureus caused 71%.19 This switch is explained
by shifting patterns of intravenous drug use. In 1983, in response to an epidemic of
pentazocine abuse, the manufacturer coformulated it with the narcotic antagonist
naloxone. With oral administration, naloxone is rapidly inactivated, but with injection,
it negates the intoxicating effect of pentazocine. Thereafter, heroin supplanted pentaz-
ocine as the drug of choice among injection drug users.63 Intravenous drug users often
inject drugs with the most convenient water supply, such as puddles in alleys or toilet
water. In one study, 81% of injection drug users reported injecting in public toilets in
the past 6 months, and 80% in the street.64 Unlike pentazocine, which dissolves in wa-
ter at room temperature, heroin must be heated to dissolve, typically by boiling in a
spoon held over a flame. This process of “cooking” reduces contamination with envi-
ronmental bacteria, such as Pseudomonas sp. Cooking does not diminish the risk of
staphylococcal infection, probably because the bacterial source is likely a colonized
skin site, rather than environmental water.

Children
In pediatric septic arthritis, knee involvement is not as dominant as in adults. The knee
and hip are infected in one-third of cases each in children.65 Staphylococci and strep-
tococci are responsible for most cases. Haemophilus influenzae type b septic arthritis
is now uncommon because of the protection provided by conjugate vaccines.66
Kingella kingae, one of the fastidious gram-negative rods of the HACEK group, has
recently been recognized as an important cause of septic arthritis, osteomyelitis, and
intervertebral diskitis in children younger than 2 years. Kingella septic arthritis is often
preceded by pharyngitis or stomatitis. Seasonality of infection has been observed,
perhaps related to viral infection or other cofactors. There is 1 report of an outbreak
of invasive Kingella infections in a day care. Routine inoculation of pediatric synovial
fluid specimens into aerobic blood culture bottles has been recommended, instead
of direct plating of specimens on solid media, to improve recovery of Kingella.67–69
Community-acquired MRSA is a major pediatric pathogen in the United States. In a
recent series from Texas, MRSA caused 62% of pediatric musculoskeletal infection.70
Disabling sequelae, such as decreased joint range of motion and limping gait, may be
more common with MRSA infection.71
 Septic Arthritis of Native Joints 211

Prosthetic Joints
The diagnosis and management of prosthetic joint infection are reviewed elsewhere in
this issue.

DIAGNOSIS: PROBLEMS AND PITFALLS

Gram stain and culture of synovial fluid should be sent in any case of undiagnosed
arthritis. Antibiotic therapy ideally should be deferred until after synovial fluid is
sampled. The involvement of rheumatology, orthopedic surgery, or interventional radi-
ology may be necessary to obtain synovial fluid.
Gram stains of synovial fluid are helpful when positive, but they are not sensitive for
the diagnosis of septic arthritis. Gram stains are positive in 71% of gram-positive sep-
tic arthritis,15 40% to 50% of cases of gram-negative septic arthritis,39–41 and fewer
than 25% of cases of gonococcal septic arthritis.2
Patients should be treated empirically for septic arthritis when synovial fluid white
blood cell (WBC) counts exceed 50,000 cells/mm3,72 although gout and pseudogout
also commonly cause WBC counts of this magnitude.73 Unfortunately, lower WBC
counts do not exclude septic arthritis. In one study, one-third of patients with septic
arthritis had synovial fluid WBCs of fewer than 50,000 cells/mm3,35 and in another
study, 50% of patients with septic arthritis had synovial fluid WBC counts fewer
than 28,000 cells/mm3.74 Immunosuppressed patients may lack synovial leukocytosis
altogether. Synovial chemistry tests, such as glucose and protein, are not useful in the
diagnosis of septic arthritis.75 Serum inflammatory markers, such as WBC counts,
erythrocyte sedimentation rates, and C-reactive protein are also of limited value, as
they are elevated in crystalline arthropathies and other conditions that may mimic sep-
tic arthritis.76 However, the serum procalcitonin level shows some promise as a
marker for septic arthritis.77,78
Blood cultures should be obtained in all patients with suspected septic arthritis. At
least one-third of patients with septic arthritis have associated bacteremia. In up to
14% of patients, a bacteriologic diagnosis is made only on the basis of blood cul-
tures.1,24 Serologic testing for Lyme disease should be obtained from patients with
undiagnosed inflammatory arthritis in endemic areas, particularly if Gram stain and
culture of synovial fluid are negative.
The diagnosis of gonococcal septic arthritis may be elusive. Fewer than 50% of sy-
novial fluid cultures are positive, even when appropriately subcultured onto chocolate
agar. The diagnosis is usually based on a clinical syndrome compatible with dissem-
inated gonococcal infection, and isolation of N gonorrhoeae from cultures or nucleic
acid amplification tests from cervical, urethral, rectal, or oropharyngeal samples.
Bacteremia is uncommon in disseminated gonococcal infection, despite the fre-
quency of polyarticular involvement.2,24,38
Approximately 20% of cases of suspected septic arthritis have negative cultures of
synovial fluid on solid media. There are many possible explanations for this phenom-
enon: the clinical diagnosis of septic arthritis may be mistaken; synovial fluid was ob-
tained after the initiation of antibiotics; small numbers of bacteria were present,
perhaps because of brisk neutrophil phagocytosis; the quantity of synovial fluid plated
was inadequate; or infecting bacteria may have had fastidious growth requirements.
Many of these problems may be overcome by inoculation of synovial fluid into blood
culture bottles. Antibiotics and other bacterial inhibitors, such as complement, may be
diluted in blood culture bottles; lytic agents present in most blood culture media, such
as saponin, may release intracellular bacteria; larger amounts of synovial fluid can be
inoculated into blood cultures; and blood cultures better support the growth of fussy
212 Ross

organisms, such as Kingella and nutritionally variant streptococci. Several studies

have shown a higher yield for pathogens, and fewer contaminants, with inoculation
of synovial fluid into blood cultures, compared with solid media.79–81
CT or MRI may confirm the presence of effusions and inflammation in joints that are
difficult to assess clinically, such as the shoulder, hip, sacroiliac, and sternoclavicular
joints (Fig. 1). It may also help to exclude osteomyelitis in adjacent bones.

ANTIBIOTIC THERAPY

Because synovial fluid tests lack precision for diagnosing septic arthritis, the threshold
for starting antibiotics should be low. Patient epidemiology may help tailor empiric
therapy toward likely organisms. However, because septic arthritis is so rapidly
destructive, broad-spectrum antibiotics are usually warranted until culture data are
available. Given the increasing importance of MRSA as a cause of septic arthritis,
initial antibiotic regimens should generally include an antibiotic active against
MRSA, such as vancomycin. Cefazolin is a reasonable alternative in areas with a
low prevalence of MRSA. A drug active against gram-negative bacilli, such as cefe-
pime or an antipseudomonal beta-lactam, should be added if patients are critically
ill or have a higher risk of gram-negative infection, such as the elderly, the immuno-
compromised, or intravenous drug users.
Septic arthritis associated with human or animal bites should be treated with agents
active against oral flora, such as ampicillin-sulbactam. Sexually active patients with
clinical syndromes suggestive of disseminated gonococcal infection should receive
ceftriaxone. Recommendations for empiric antibiotic therapy of suspected septic
arthritis are summarized in Box 2.
Once the causative organism has been identified, therapy should be narrowed
based on sensitivity data. Data on duration of therapy are scanty. In general, septic
arthritis in adults should be treated for at least 3 weeks, which may include a period
of step-down oral therapy. In children with uncomplicated septic arthritis, as few as
10 days of antibiotic therapy may suffice.82 Gonococcal septic arthritis can be treated
with 2 weeks of ceftriaxone. Because osteomyelitis is common in infections of carti-
laginous joints, such as the sternoclavicular and sacroiliac joints, antibiotic courses
of 6 weeks are recommended.23

Fig. 1. A 35-year-old man with a history of intravenous drug use presented with a painful
left hip. Pelvic MRI (coronal cut) showed a moderate effusion of the left hip joint (arrow),
with diffuse edema involving nearly all of the muscles of the left hip. Blood cultures and sy-
novial fluid cultures grew methicillin-sensitive S aureus.
 Septic Arthritis of Native Joints 213

Box 2
Empiric antibiotic therapy for suspected septic arthritis of native joints

Gram-positive cocci on Gram stain of synovial fluid

High-prevalence area for methicillin-resistant Staphylococcus aureus (MRSA)
Vancomycin 1 g intravenous (IV) every 12 hours
Low-prevalence area for MRSA
Cefazolin 2 g IV every 8 hours
Gram-negative cocci on Gram stain of synovial fluid, or clinical syndrome suggestive of
disseminated gonococcal infection
Ceftriaxone 1 g IV every 24 hours PLUS
Azithromycin 1 g orally (single dose)
Gram-negative rods on Gram stain of synovial fluid
Cefepime 2 g IV every 8 hours or piperacillin-tazobactam 4.5 g IV every 6 hours
No organisms seen on Gram stain
Vancomycin 1 g IV every 12 hours (cefazolin may be used in areas of low MRSA prevalence)
In the elderly, immunocompromised, critically ill, or intravenous drug users, add an
antipseudomonal beta-lactam, such as cefepime or piperacillin-tazobactam

Historical series of septic arthritis have shown that a large proportion of patients
with suspected septic arthritis have negative joint fluid cultures. These patients have
similar clinical and epidemiologic features compared with those with bacteriologically
confirmed septic arthritis. It is reasonable to complete a short course of oral antibiotic
therapy in these patients.83

JOINT DRAINAGE

In the preantibiotic era, some patients with septic arthritis had good outcomes with
aggressive joint irrigation alone. Today, septic arthritis is managed with antibiotics
combined with joint drainage by arthroscopy, arthrocentesis, or arthrotomy. Joint
drainage decompresses the joint, improves blood flow, and removes bacteria, toxins,
and proteases. Arthrocentesis should be repeated daily until effusions resolve and
cultures are negative. There is general agreement that surgical drainage is indicated
for septic arthritis of the hip, failure to respond after 5 to 7 days of antibiotics
and arthrocentesis, and soft tissue extension of infection. The shoulder joint should
be drained either surgically or under radiologic guidance.84,85 Retrospective data
suggest that patients with RA have better functional outcomes with surgical
management.9
Patients with sternoclavicular septic arthritis often respond poorly to medical man-
agement, especially when antibiotic therapy is delayed. Thoracic surgery should be
consulted in these patients to assess the need for sternoclavicular joint excision
and pectoralis flap grafting.86,87 Chest imaging with CT or MRI should be obtained
routinely to exclude osteomyelitis, chest wall abscess, and mediastinitis.
No good data show a superiority of surgical drainage over arthrocentesis. In fact,
one meta-analysis, and a more recent retrospective study, demonstrated better func-
tional outcomes with arthrocentesis compared with surgery, although mortality was
higher in patients treated with arthrocentesis.13,88 Selection bias probably explains
these differences. Critically ill patients are poor surgical candidates, whereas other-
wise stable patients with severe septic arthritis are more likely to undergo surgical
drainage. Randomized clinical trials of arthrocentesis compared with surgical or
arthroscopic drainage are needed.
214 Ross

A recent study showed a significant benefit of dexamethasone therapy in preventing

disability in children with septic arthritis, but no data exist to recommend its use in
adults.89
Aggressive rehabilitation is essential to prevent joint contractures and muscle atro-
phy. Patients should be mobilized as soon as pain allows. Although early reviews rec-
ommended joint immobilization in the acute phase of infection,2 in animal models of
septic arthritis, more cartilage degeneration and adhesions were seen in immobilized
animals, compared with animals treated with continuous passive motion devices.90

REFERENCES

1. Cooper C, Cawley MI. Bacterial arthritis in an English health district: a 10 year re-
view. Ann Rheum Dis 1986;45:458–63.
2. Goldenberg DL, Reed JI. Bacterial arthritis. N Engl J Med 1985;312:764–71.
3. Kaandorp CJ, Dinant HJ, van de Laar MA, et al. Incidence and sources of native
and prosthetic joint infection: a community based prospective survey. Ann
Rheum Dis 1997;56:470–5.
4. Haug JB, Harthug S, Kalager T, et al. Bloodstream infections at a Norwegian uni-
versity hospital, 1974–1979 and 1988–89: changing etiology, clinical features,
and outcome. Clin Infect Dis 1994;19:246–56.
5. Ross JJ, Saltzman CL, Carling P, et al. Pneumococcal septic arthritis: review of
190 cases. Clin Infect Dis 2003;36:319–27.
6. Patti JM, Bremell T, Krajewska-Pietrasik D, et al. The Staphylococcus aureus
collagen adhesin is a virulence determinant in experimental septic arthritis. Infect
Immun 1994;62:152–61.
7. Shirtliff ME, Mader JT. Acute septic arthritis. Clin Microbiol Rev 2002;15:527–44.
8. Stevens CR, Williams RB, Farrell AJ, et al. Hypoxia and inflammatory synovitis:
observations and speculation. Ann Rheum Dis 1991;50:124–32.
9. Gardner GC, Weisman MH. Pyarthrosis in patients with rheumatoid arthritis: a
report of 13 cases and a review of the literature from the past 40 years. Am J
Med 1990;88:503–11.
10. Ilahi OA, Swarna U, Hamill RJ, et al. Concomitant crystal and septic arthritis. Or-
thopedics 1996;19:613–7.
11. Yu KH, Luo SF, Liou LB, et al. Concomitant septic and gouty arthritis: an analysis
of 30 cases. Rheumatology (Oxford) 2003;42:1062–6.
12. Galloway JB, Hyrich KL, Mercer LK, et al. Risk of septic arthritis in patients with
rheumatoid arthritis and the effect of anti-TNF therapy: results from the British So-
ciety for Rheumatology biologics register. Ann Rheum Dis 2011;70:1810–4.
13. Weston VC, Jones AC, Bradbury N, et al. Clinical features and outcome of septic
arthritis in a single UK Health District 1982–1991. Ann Rheum Dis 1999;58:214–9.
14. Ross JJ. Septic arthritis. Infect Dis Clin North Am 2005;19:799–817.
15. Goldenberg DL, Cohen AS. Acute infectious arthritis. Am J Med 1976;60:369–77.
16. Edwards SA, Cranfield T, Clarke HJ. Atypical presentation of septic arthritis in the
immunosuppressed patient. Orthopedics 2002;25:1089–90.
17. Jones A, Doherty M. ABC of rheumatology: osteoarthritis. BMJ 1995;310:457–60.
18. Dubost JJ, Fis I, Denis P, et al. Polyarticular septic arthritis. Medicine (Baltimore)
1993;72:296–310.
19. Brancos MA, Peris P, Miro JM, et al. Septic arthritis in heroin addicts. Semin
Arthritis Rheum 1991;21:81–7.
 Septic Arthritis of Native Joints 215

20. Vyskocil JJ, McIlroy MA, Brennan TA, et al. Pyogenic infection of the sacroiliac
joint: case reports and review of the literature. Medicine (Baltimore) 1991;70:
188–97.
21. Zimmermann B III, Mikolich DJ, Lally EV. Septic sacroiliitis. Semin Arthritis Rheum
1996;26:592–604.
22. Ross JJ, Hu LT. Septic arthritis of the pubic symphysis: review of 100 cases. Med-
icine (Baltimore) 2003;82:340–5.
23. Ross JJ, Shamsuddin H. Sternoclavicular septic arthritis: review of 180 cases.
Medicine (Baltimore) 2004;83:139–48.
24. Sharp JT, Lidsky MD, Duffy J, et al. Infectious arthritis. Arch Intern Med 1979;139:
1125–30.
25. Ross JJ, Davidson L. Methicillin-resistant Staphylococcus aureus septic arthritis:
an emerging clinical syndrome. Rheumatology (Oxford) 2005;44:1197–8.
26. Daynes J, Roth MF, Zekaj M, et al. Adult native septic arthritis in an inner city hos-
pital: effects on length of stay. Orthopedics 2016;39:e674–9.
27. Frazee BW, Fee C, Lambert L. How common is MRSA in adult septic arthritis?
Ann Emerg Med 2009;54:695–700.
28. Helito CP, Noffs GG, Pecora JR, et al. Epidemiology of septic arthritis of the knee
at Hospital das Clı́nicas, Universidade de São Paulo. Braz J Infect Dis 2014;18:
28–33.
29. Chao CM, Lai CC, Hsueh PR. Bacteriology of septic arthritis at a regional hospital
in southern Taiwan. J Microbiol Immunol Infect 2013;46:241–2.
30. Dubost JJ, Couderc M, Tatar Z, et al. Three-decade trend in the distribution of or-
ganisms causing septic arthritis in native joints. Joint Bone Spine 2014;81(5):
438–40.
31. Clerc O, Prod’hom G, Greub G, et al. Adult native septic arthritis: a review of 10
years of experience and lessons for empirical antibiotic therapy. J Antimicrob
Chemother 2011;66:1168–73.
32. Farley MM. Group B streptococcal disease in nonpregnant adults. Clin Infect Dis
2001;33:556–61.
33. Dubost JJ, Soubrier M, De Champs C, et al. No changes in the distribution of or-
ganisms responsible for septic arthritis over a 20 year period. Ann Rheum Dis
2002;61:267–9.
34. Nolla JM, Gomez-Vaquero C, Corbella X, et al. Group B streptococcus (Strepto-
coccus agalactiae) pyogenic arthritis in nonpregnant adults. Medicine (Balti-
more) 2003;82:119–27.
35. Li SF, Henderson J, Dickman E, et al. Laboratory tests in adults with monoarticular
arthritis: can they rule out a septic joint? Acad Emerg Med 2004;11:276–80.
36. Belkhir L, Rodriguez-Villalobos H, Vandercam B, et al. Pneumococcal septic
arthritis in adults: clinical analysis and review. Acta Clin Belg 2014;69:40–6.
37. Pulido M, Sorvillo F. Declining invasive pneumococcal disease mortality in the
United States, 1990-2005. Vaccine 2010;28:889–92.
38. O’Brien JP, Goldenberg DL, Rice PA. Disseminated gonococcal infection: a pro-
spective analysis of 49 patients and a review of pathophysiology and immune
mechanisms. Medicine (Baltimore) 1983;62:395–406.
39. Goldenberg DL, Brandt KD, Cathcart ES, et al. Acute arthritis due to gram-
negative bacilli: a clinical characterization. Medicine (Baltimore) 1974;53:
197–208.
40. Bayer AS, Chow AW, Louie JS, et al. Gram-negative bacillary septic arthritis: clin-
ical, radiographic, therapeutic, and prognostic features. Semin Arthritis Rheum
1977;7:123–32.
216 Ross

41. Newman ED, Davis DE, Harrington TM. Septic arthritis due to gram negative
bacilli: older patients with good outcome. J Rheumatol 1988;15:659–62.
42. Schaad UB. Arthritis in disease due to Neisseria meningitidis. Rev Infect Dis
1980;2:880–8.
43. Goedvolk CA, von Rosenstiel IA, Bos AP. Immune complex associated complica-
tions in the subacute phase of meningococcal disease: incidence and literature
review. Arch Dis Child 2003;88:927–30.
44. Rompalo AM, Hook EW III, Roberts PL, et al. The acute arthritis-dermatitis syn-
drome: the changing importance of Neisseria gonorrhoeae and Neisseria menin-
gitidis. Arch Intern Med 1987;147:281–3.
45. Armstrong RW, Bolding F, Joseph R. Septic arthritis following arthroscopy: clinical
syndromes and analysis of risk factors. Arthroscopy 1992;8:213–23.
46. Indelli PF, Dillingham M, Fanton G, et al. Septic arthritis in postoperative anterior
cruciate ligament reconstruction. Clin Orthop 2002;398:182–8.
47. Gardam M, Lim S. Mycobacterial osteomyelitis and arthritis. Infect Dis Clin North
Am 2005;19:819–30.
48. Eder L, Zisman D, Rozenbaum M, et al. Clinical features and aetiology of septic
arthritis in northern Israel. Rheumatology (Oxford) 2005;44:1559–63.
49. Gotuzzo E, Alarcon GS, Bocanegra TS, et al. Articular involvement in human
brucellosis: a retrospective analysis of 304 cases. Semin Arthritis Rheum 1982;
12:245–55.
50. Solera J, Rodriguez-Zapata M, Geijo P, et al. Doxycycline-rifampin versus
doxycycline-streptomycin in treatment of human brucellosis due to Brucella me-
litensis. Antimicrob Agents Chemother 1995;39:2061–7.
51. Kratz A, Greenberg D, Barki Y, et al. Pantoea agglomerans as a cause of septic
arthritis after palm tree thorn injury; case report and literature review. Arch Dis
Child 2003;88:542–4.
52. Freiberg AA, Herzenberg JE, Sangeorzan JA. Thorn synovitis of the knee joint
with Nocardia pyarthrosis. Clin Orthop 1993;287:233–6.
53. Talan DA, Abrahamian FM, Moran GJ, et al. Clinical presentation and bacterio-
logic analysis of infected human bites in patients presenting to emergency de-
partments. Clin Infect Dis 2003;37:1481–9.
54. Talan DA, Citron DM, Abrahamian FM, et al. Bacteriologic analysis of infected
dog and cat bites. N Engl J Med 1999;340:85–92.
55. Chevalier X, Martigny J, Avouac B, et al. Report of 4 cases of Pasteurella multo-
cida septic arthritis. J Rheumatol 1991;18:1890–2.
56. Feurle GE, Marth T. An evaluation of antimicrobial treatment for Whipple’s dis-
ease: tetracycline versus trimethoprim-sulfamethoxazole. Dig Dis Sci 1994;39:
1642–8.
57. Mahnel R, Kalt A, Ring S, et al. Immunosuppressive therapy in Whipple’s disease
patients is associated with the appearance of gastrointestinal manifestations. Am
J Gastroenterol 2005;100:1167–73.
58. Lange U, Teichmann J. Whipple arthritis: diagnosis by molecular analysis of sy-
novial fluid–current status of diagnosis and therapy. Rheumatology (Oxford)
2003;42:473–80.
59. Mahnel R, Marth T. Progress, problems, and perspectives in diagnosis and treat-
ment of Whipple’s disease. Clin Exp Med 2004;4:39–43.
60. Asmar BI, Andresen J, Brown WJ. Ureaplasma urealyticum arthritis and bacter-
emia in agammaglobulinemia. Pediatr Infect Dis J 1998;17:73–6.
 Septic Arthritis of Native Joints 217

61. O’Sullivan MV, Isbel NM, Johnson DW, et al. Disseminated pyogenic Mycoplasma
pneumoniae infection in a renal transplant recipient, detected by broad-range
polymerase chain reaction. Clin Infect Dis 2004;39:e98–9.
62. Mian AN, Farney AC, Mendley SR. Mycoplasma hominis septic arthritis in a pe-
diatric renal transplant recipient: case report and review of the literature. Am J
Transplant 2005;5:183–8.
63. Baum C, Hsu JP, Nelson RC. The impact of the addition of naloxone on the use
and abuse of pentazocine. Public Health Rep 1987;102:426–9.
64. Darke S, Kaye S, Ross J. Geographic injecting locations among injecting drug
users in Sydney, Australia. Addiction 2001;96:241–6.
65. Krogstad P, Smith AL. Osteomyelitis and septic arthritis. In: Feigin RD, Cherry JD,
editors. Textbook of pediatric infectious disease. 4th edition. Philadelphia: WB
Saunders; 1998. p. 683–704.
66. Luhmann JD, Luhmann SJ. Etiology of septic arthritis in children: an update for
the 1990s. Pediatr Emerg Care 1999;15:40–2.
67. Yagupsky P, Peled N, Katz O. Epidemiological features of invasive Kingella
kingae infections and respiratory carriage of the organism. J Clin Microbiol
2002;40:4180–4.
68. Centers for Disease Control and Prevention. Osteomyelitis/septic arthritis caused
by Kingella kingae among day care attendees in Minnesota, 2003. MMWR Morb
Mortal Wkly Rep 2004;53:241–3.
69. Centers for Disease Control and Prevention. Kingella kingae infections in chil-
dren. United States, June 2001-November 2002. MMWR Morb Mortal Wkly Rep
2004;53:244.
70. Erickson CM, Sue PK, Stewart K, et al. Sequential parenteral to oral clindamycin
dosing in pediatric musculoskeletal infection. Pediatr Infect Dis J 2016;35:
1092–6.
71. Wang CL, Wang SM, Yang YJ, et al. Septic arthritis in children: relationship of
causative pathogens, complications, and outcome. J Microbiol Immunol Infect
2003;36:41–6.
72. Margaretten ME, Kohlwes J, Moore D, et al. Does this adult patient have septic
arthritis? JAMA 2007;297:1478–88.
73. Fye KH. Arthrocentesis, synovial fluid analysis, and synovial biopsy. In:
Klippel JH, editor. Primer on the rheumatic diseases. 12th edition. Atlanta:
Arthritis Foundation; 2001. p. 138–44.
74. McCutchan HJ, Fisher RC. Synovial leukocytosis in infectious arthritis. Clin Or-
thop 1990;257:226–30.
75. Shmerling RH, Delbanco TL, Tosteson AN, et al. Synovial fluid tests. What should
be ordered? JAMA 1990;264:1009–14.
76. Carpenter CR, Schuur JD, Everett WW, et al. Evidence-based diagnostics: adult
septic arthritis. Acad Emerg Med 2011;18:781–96.
77. Shen CJ, Wu MS, Lin KH, et al. The use of procalcitonin in the diagnosis of bone
and joint infections: a systematic review and meta-analysis. Eur J Clin Microbiol
Infect Dis 2013;32:807–14.
78. Paosong S, Narongroeknawin P, Pakchotanon R, et al. Serum procalcitonin as a
diagnostic aid in patients with acute bacterial septic arthritis. Int J Rheum Dis
2015;18:353–9.
79. Yagupsky P, Press J. Use of the isolator 1.5 microbial tube for culture of synovial
fluid from patients with septic arthritis. J Clin Microbiol 1997;35:2410–2.
218 Ross

80. Hughes JG, Vetter EA, Patel R, et al. Culture with BACTEC Peds Plus/F bottle
compared with conventional methods for detection of bacteria in synovial fluid.
J Clin Microbiol 2001;39:4468–71.
81. Hepburn MJ, Fraser SL, Rennie TA, et al. Septic arthritis caused by Granulicatella
adiacens: diagnosis by inoculation of synovial fluid into blood culture bottles.
Rheumatol Int 2003;23:255–7.
82. Pääkkönen M, Peltola H. Treatment of acute septic arthritis. Pediatr Infect Dis J
2013 Jun;32(6):684–5.
83. Gupta MN, Sturrock RD, Field M. Prospective comparative study of patients with
culture proven and high suspicion of adult onset septic arthritis. Ann Rheum Dis
2003;62:327–31.
84. Pioro MH, Mandell BF. Septic arthritis. Rheum Dis Clin North Am 1997;23:239–58.
85. Smith JW, Piercy EA. Infectious arthritis. Clin Infect Dis 1995;20:225–31.
86. Song HK, Guy TS, Kaiser LR, et al. Current presentation and optimal surgical
management of sternoclavicular joint infection. Ann Thorac Surg 2002;73:427–31.
87. Kachala SS, D’Souza DM, Teixeira-Johnson L, et al. Surgical management of ster-
noclavicular joint infections. Ann Thorac Surg 2016;101:2155–60.
88. Broy SB, Schmid FR. A comparison of medical drainage (needle aspiration) and
surgical drainage (arthrotomy or arthroscopy) in the initial treatment of infected
joints. Clin Rheum Dis 1986;12:501–22.
89. Odio CM, Ramirez T, Arias G, et al. Double blind, randomized, placebo-controlled
study of dexamethasone therapy for hematogenous septic arthritis in children.
Pediatr Infect Dis J 2003;22:883–9.
90. Salter RB. The biologic concept of continuous passive motion of synovial joints:
the first 18 years of research and its clinical application. Clin Orthop 1989;242:
12–25.