Patients with the Systemic Inflammatory Response Syndrome (SIRS) following trauma may or may not be infected. The presence of certain findings can help identify patients who have become infected. Identification of infected patients can help reduce inappropriate use of antibiotics and risk from infection caused by antibiotic-resistant organisms. The authors are from Wake Forest University in Winston-Salem, North Carolina.
Selection criteria: 59 patients in entire group, with 35 developing SIRS:
(1) patients in trauma ICU for > 48 hours
(2) patients not transferred from another institution with the diagnosis of infection
Features associated with infection in patients with the Systemic Inflammatory Response Syndrome:
(1) maximum temperature > 102°F after post-injury day 4
(2) C-reactive protein > 17 mg/dL
The WBC count was found not to be useful in determining the presence of infection.
Most of the infections were pneumonias (61%), while others had line sepsis, wound infections, intra-abdominal infection or bacteremia.
Performance of criteria:
• reported performance (see NOTE): sensitivity 91%, specificity 50% (for SIRS group)
• C reactive protein alone: sensitivity 74%, specificity 75% (for all patients studied); sensitivity 72% and specificity 100% (for SIRS group)
• temperature alone: sensitivity 67%, specificity 80% (for all patients studied); sensitivity 72% and specificity 83% (for SIRS group)
NOTE: I am confused by the claim for sensitivity of 91% and specificity of 50%. How can both tests together be more sensitive than either test alone? My personal belief is that the sensitivity and specificity need to be reversed. Having a 2x2 table in the article would have been nice. I tried several combinations of findings and could not find a reasonable data set to explain the reported sensitivity and specificity for the SIRS patients; the stated performance could have been for the entire study population.
How well this protocol would apply to nontraumatic patients with SIRS is unclear.
The criterion to wait for more than 4 days after injury reflects the length of time for infection to become established after the injury.
