Note: "Septic arthritis" in this post and discussion refers to non-gonococcal septic arthritis.
What's the big deal with septic arthritis? It's not really an emergency is it?
Within days of onset, septic arthritis destroys cartilage and can leave the joint with permanent joint disability occurring in 25-50% of the cases.
The mortality rate for in-hospital septic arthritis ranges from 7% to 15%, despite antibiotic use.
Ok ok. So that's pretty bad. How do I make sure I don't miss a septic arthritis?
Unfortunately, there are no physical examination findings or maneuvers, such as range of motion or degree of swelling, that have been studied that can help the clinician discriminate between etiologies of the monoarthritis.
Basically in any patient presenting with a single painful joint, there is no way to rule out a septic arthritis without an arthrocentesis and synovial culture.
That said, even synovial culture is not a perfect test.
Granted, we don't tap every painful joint, because we often have an alternative diagnosis that results in a very low pre-test probability for a septic joint, and other risk factors or lack-of risk factors lowers the probability of disease even more. However, just keep in mind that you can never exclude a septic joint in a painful joint completely by physical exam and history alone. It can only be truly excluded with an arthrocentesis.
Below are some risk factors/likelihood ratios to definitely raise your suspicion for septic arthritis (1). But remember, that just because they have zero risk factors, doesn’t mean that they won’t have a septic joint. It can happen to anyone.
Age >80 (+LR 3.5)
Diabetes (+LR 2.7)
Rheumatoid arthritis (+LR 3.5)
Joint surgery within past 3 mo (+LR 6.9)
Hip/knee prosthesis (+LR 3.1)
Skin infection (+LR 2.8)
Skin infection + prosthesis (+LR 15)
List of common myths and dogma regarding septic arthritis:
“The patient doesn’t have a fever, so it’s not a septic joint”
Fever Sensitivity 46%. Specifificity 31% (1)
So you’d only miss 54% of septic joints relying on fever.
"If the serum WBC isn't elevated, it's not a septic joint."
Sensitivity WBC >10,000 is 90% (1)
If you're ok with missing 10% of septic joints.
"If the CRP and ESR aren't elevated, it's not a septic joint"
Sensitivity ESR >30 is 95% (1) - miss 5% of septic joints relying on a very low cutoff for ESR
Sensitivity CRP >100 is 77% (1) - miss 23% of septic joints relying on CRP
"If the synovial WBC <50,000 it's not a septic joint” (My personal favorite)
Sensitivity WBC >50,000 is 62% (1)
So 38% of septic joints are missed relying on a synovial WBC >50,000
"If the neutrophile count is <90% it's not a septic joint"
Sensitivity PMN >90% is 73% (1)
Missing 27% of septic joints
"If the gram stain is negative, it's not a septic joint"
Sensitivity of gram stain 50-80% (2)
Missing 20-50% of septic joints.
You have to follow the culture for 48 hours.
Depending on their risk factors, your clinical suspicion, and synovial WBC, you may or may not start antibiotics until the cultures come back
"If there are crystals, it's not a septic joint"
5% of crystal monoarthritis have a concomitant septic arthritis (3)
Crystals have also been found in asymptomatic patients, so crystals definitely can’t exclude an infection
"Synovial fluid culture is the gold standard rule out test for septic arthritis. If the culture is negative, it's not septic arthritis."
Synovial fluid culture is the most important test we have to rule out septic arthritis, but it's not perfect. There have been multiple reports of cases of septic arthritis where other diagnostic tests (such as imaging), clinical course, etc resulted in a definitive diagnosis of septic arthritis despite an initial negative culture from synovial fluid. (4, 5)
The incidence of culture-negative septic arthritis is certainly small, but incredibly difficult to give a number too since there is no true gold-standard test to compare a 48 hour synovial culture to.
The take home message is that when septic arthritis is suspected, empiric antimicrobial therapy is warranted following arthrocentesis until culture data are available. Even if the gram stain and culture are negative, if the patient seems to be responding to empiric therapy, continuing a full treatment course of antibiotics may be necessary given the lack of any true gold standard diagnostic test.
How do I perform an arthrocentesis?
Consider using ultrasound. (I’m a big fan of ultrasound as an adjunct to look for the effusion… especially for small joints like wrists and ankles.)
18 or 20 gauge needle to aspirate fluid
Location of needle entry for ankle, elbow, knee, shoulder via very photo from @ClinicalEMed below: