Background: Systemic inflammatory response syndrome (SIRS) is common in trauma patients, and infection represents an important and treatable source of SIRS. C-Reactive protein (CRP), an acute phase protein, is elevated in infection and discriminates between infected and uninfected patients in other patient populations. Our goal was to examine the ability of CRP and other commonly used markers of infection (maximum temperature [Tmax], and white blood cell count [WBC]) to distinguish between infectious and noninfectious causes of SIRS.
Methods: This was a prospective study of a consecutive series of trauma patients who spent greater than 48 hours in the intensive care unit. Studied variables included CRP, Tmax, WBC, and culture-proven infection compared with standard definitions of infection and the presence of SIRS. The ability of these variables to correctly classify patients as infected (INF) or not infected was examined by using receiver operating characteristic curves. Values on the day of infection diagnosis in the INF group and on postadmission day 5 (the mean day of onset of infection in the INF group) in the not infected group were used. Multivariate discriminant analysis was used to examine the relative contributions of Tmax and CRP in predicting infection. Significance was defined as p < 0.05.
Results: Fifty-nine patients were admitted over a 4-month period. Of these, 35 patients (59%) had SIRS at the time of comparison (29 INF, 6 not infected). Thirty-three patients (56%) developed an infection. Both CRP and Tmax discriminated between patients with and without infection whereas WBC did not (areas under receiver operating characteristic curve: 0.86, 0.81, and 0.47, respectively). In patients with SIRS, cutoff values of 17 mg/dL for CRP (specificity 100%) and 102 degrees F for Tmax (specificity 83%) were identified. CRP added significant discriminatory power to Tmax in determining presence of infection in patients with SIRS (p = 0.003).
Conclusion: Infection must be presumed to be the source of SIRS in patients with CRP more than 17 mg/dL and Tmax more than 102 degrees F after postinjury day 4. WBC is not useful in determining the presence of infection.