Does this patient have a septic joint?
The difficulty in diagnosing this elusive disease is that the history, physical, and serum tests are typically unhelpful in ruling in or out the disease. See my previous Paucis Verbis card covering the 2007 JAMA review on Septic Arthritis.
Interpretation of the synovial fluid is also challenging. Generally a synovial WBC count <25K suggests a non-infectious process and a WBC count >100K suggests an infectious process.
What about everything in between 25K and 100K?
Trick of the Trade:
Check if the synovial lactate level is > 10 mmol/L.
A level >10 mmol/L is highly suggestive of septic arthritis. The calculated positive likelihood ratio (LR) from multiple studies was at least 10. The negative LR ranged from 0 to 0.45.
Although the quoted studies measured lactate using liquid chromatography, presumably our point-of-care lactate levels are equivalent, based on various sepsis studies.
How do you use the LR statistic? I'm no statistician, and so I love the Bayes nomogram. This requires me to have a pretest probability for the disease. In this case, let's say that I am moderately suspicious of a septic joint given the patient's history of rheumatoid arthritis (a known risk factor), significant joint pain, and low grade fever. I'm going to say that my pretest probability is 25% (see left column of numbers).
The synovial lactate level returns at 12 mmol/L (see middle column of numbers), which gives the patient a positive LR of at least 10.
This means that my post-test probability for a septic joint jumps way up to 80%, which practically rules-in my patient for septic joint.
Reference
Carpenter CR, Schuur JD, Everett WW, Pines JM. Evidence-based Diagnostics: Adult Septic Arthritis. Acad Emerg Med. 2011;18(8), 781-96. PMID: 21843213
.
