Does SOFA predict outcomes better than SIRS in Brazilian ICU patients with suspected infection? A retrospective cohort study

Does SOFA predict outcomes better than SIRS in Brazilian ICU patients with suspected infection? A retrospective cohort study

Brazilian Journal of Infectious Diseases, October 2017

29035701

Regis Goulart Rosa, Rafael Barberena Moraes, Thiago Costa Lisboa, Daniel Pretto Schunemann, Cassiano Teixeira

We compared the discriminatory capacity of the sequential organ failure assessment (SOFA) versus the systemic inflammatory response syndrome (SIRS) score for predicting ICU mortality, need for and length of mechanical ventilation (MV), ICU stay, and hospitalization in patients with suspected infection admitted to a mixed Brazilian ICU. We performed a retrospective analysis of a longitudinal ICU database from a tertiary hospital in Southern Brazil. Patients were categorized according to whether they met the criteria for sepsis according to SOFA (variation ≥2 points over the baseline clinical condition) and SIRS (SIRS score ≥2 points). From January 2008 to December 2014, 1487 patients were admitted to the ICU due to suspected infection. SOFA ≥2 identified more septic patients than SIRS ≥2 (79.0% [n=1175] vs. 68.5% [n=1020], p<0.001). There was no difference between the two scores in predicting ICU mortality (area under the receiver operating characteristic curve (AUROC)=0.64 vs. 0.64, p=0.99). SOFA ≥2 was marginally better than SIRS ≥2 in predicting need for MV (AUROC=0.64 vs. 0.62, p=0.001), ICU stay>7 days (AUROC=0.65 vs. 0.63, p=0.004), and length of hospitalization >10 days (AUROC=0.61 vs. 0.59, p<0.001). There was no difference between the two scores in predicting MV >7 days.