Modified criteria for the systemic inflammatory response syndrome improves their utility following cardiac surgery

Abstract

Background: Debate remains regarding whether the systemic inflammatory response syndrome (SIRS) identifies patients with clinically important inflammation. Defining criteria may be disproportionately sensitive and lack specificity. We investigated the incidence and evolution of SIRS in a homogenous population (following cardiac surgery) over 7 days to establish the relationship between SIRS and outcome, modeling alternative permutations of the criteria to increase their discriminatory power for mortality, length of stay, and organ dysfunction.

Methods: We conducted a retrospective analysis of prospectively collected data from a cardiothoracic ICU. Consecutive patients requiring ICU admission for the first time after cardiac surgery (N = 2,764) admitted over a 41-month period were studied.

Results: Concurrently, 96.2% of patients met the standard two criterion definition for SIRS within 24 h of ICU admission. Their mortality was 2.78%. By contrast, three or four criteria were more discriminatory of patients with higher mortality (4.21% and 10.2%, respectively). A test dataset suggested that meeting two criteria for at least 6 consecutive h may be the best model. This had a positive and negative predictive value of 7% and 99.5%, respectively, in a validation dataset. It performed well at predicting organ dysfunction and prolonged ICU admission.

Conclusions: The concept of SIRS remains valid following cardiac surgery. With suitable modification, its specificity can be improved significantly. We propose that meeting two or more defining criteria for 6 h could be used to define better populations with more difficult clinical courses following cardiac surgery. This group may merit a different clinical approach.