Diagnostic Accuracy of Synovial and Serum Markers for Septic Arthritis
November 2007
November 2007
Diagnostic Accuracy of Synovial and Serum Markers for Septic Arthritis
Search Strategy: Using PUBMED Clinical Queries you perform the following search: (“infectious arthritis”[Text Word] OR “arthritis, infectious”[MeSH Terms] OR septic arthritis [Text Word]) AND specificity [Title/Abstract] you locate three recent reviews on this topic. Reviewing the references for each of these reviews, you note additional primary studies addressing this question and pull one.
A 75-year old male presents with 2-days of excruciating left knee pain. His past medical history includes hypertension and GERD. He notes no history of prior knee injuries or surgeries. He has no history of crystalloid arthropathy or connective tissue disease. He leads an exceptionally active life style, playing tennis with his son-in-law twice per week year round. In fact, he is a retired professional tennis player and recently won his age-division’s annual tournament at the local health club. Tylenol, his typical analgesic-of-choice, has been ineffective for his knee pain. He adamantly denies any injury to the knee playing tennis or otherwise. Although he has never been to an Emergency Department for anything, the pain was just too excruciating to wait for the appointment offered by his PCP in three weeks.
On physical exam you note, hypertension (190/110) and no fever (36.7 C), sinus rhythm, clear lungs, and a red, swollen left knee measuring about 1.5x the size of the asymptomatic right knee. He has moderate pain with passive range-of-motion and you suspect a suprapatellar and infrapatellar effusion. His ligaments appear stable on a limited exam (due to pain) compared with the right side. The chest radiograph displays pulmonary venous congestion.
While discussing the differential diagnosis of an acutely erythematous, swollen, painful single joint with the patient you wonder what the diagnostic utility of various laboratory tests are in distinguishing septic arthritis from crystalloid or non-crystalloid/non-septic arthropathy.
PICO Question
Population: ED patients presenting suspected septic arthritis
Intervention: Diagnostic testing with serum WBC, synovial WBC/gram stain/crystal analysis
Comparison: Clinical gestalt alone without ancillary testing
Outcome: Diagnostic accuracy, morbidity, mortality of delayed/missed diagnosis, false-positive rates and related sequelae.
Years
First years: Diagnostic Utility of Laboratory Tests in Septic Arthritis Emerg Med J 2007; 24: 75-7. (http://pmid.us/17251607)
Second years: Does this Adult Patient Have Septic Arthritis? JAMA 2007; 297: 1497-1488. (http://pmid.us/17405973)
Articles
Article 1: Diagnostic Utility of Laboratory Tests in Septic Arthritis Emerg Med J 2007; 24: 75-77
ANSWER KEY
Article 2: Does this Adult Patient Have Septic Arthritis? JAMA 2007; 297: 1497-1488
ANSWER KEY
Article 3: Does the Presence of Crystal Arthritis Rule Out Septic Arthritis? J Emerg Med 2007; 32: 23-26
ANSWER KEY
Article 4: Bacterial or Crystal-associated Arthritis? Discriminating Ability of Serum Inflammatory Markers, Scand J Infect Dis 1998; 30: 591-596
ANSWER KEY
Bottom Line
History and physical exam, CBC, ESR, and CRP are useless for the diagnosis of septic arthritis since they do not substantially change the pre-test in patients with acutely painful, swollen joint. On the other hand, the presence of elevated synovial WBC (> 100,000 LR+ = 28, >50,000 LR+ = 7.7) can change pre-test probability sufficiently to cross the treatment threshold and modify management. Septic arthritis rarely occurs concurrently with crystalloid arthritis (1.5% prevalence) and when it does one single-center study suggested that synovial WBC > 50,000 identified all the cases of simultaneous crystalloid-septic arthritis with a NNT = 9 to treat all concurrent cases pending culture results if this synovial WBC threshold is used. Synovial fluid gram stain only demonstrates bacteria in 42% of septic arthritis cases. The prevalence of septic arthritis in ED-populations presenting with acutely painful, swollen monoarticular arthritis is between 10-28% with the 28% estimate derived from the only ED-based prospective study.