qSOFA has been proposed as a prognostic tool in patients with sepsis. This study set out to assess the sensitivity of several scores, namely: the pre-ICU qSOFA, the qSOFA with lactate (qSOFA L), SIRS score, qSOFA + SIRS score (qSIRS) and qSIRS with lactate (qSIRS L) in predicting in-hospital mortality in patients with surgical sepsis as well as the sensitivity of these scores in predicting high-grade sepsis. The secondary aim was to determine which of these scores is best suited to predict high-grade surgical sepsis.
This was a retrospective cohort study that was conducted between December 2012 and August 2017 in a public metropolitan surgical service. Data from patients aged > 13 years, who were admitted to the hospital and who had an emergency surgical procedure for source control were retrieved from a prospectively maintained hybrid electronic database. The qSOFA, qSOFA plus lactate (qSOFA L), SIRS and qSOFA + SIRS (qSIRS), as well as the qSIRS plus lactate (qSIRS L), were calculated for each patient. A lactate level that was greater than 2mmol/L was deemed to be a positive finding. Any score ≥2 was deemed to be a positive score. The outcome measure was in-hospital mortality. The prognostic value of qSOFA, qSOFA L, SIRS, qSIRS and qSIRS L was studied. Receiver operating characteristic analyses were performed to determine the area under the curve (AUC), sensitivity, specificity and positive and negative likelihood ratios for positive qSOFA, qSOFA L, SIRS, qSIRS, and qSIRS L. Contingency tables were used to calculate the sensitivity, specificity, PPV and NPV for predicting severe or high-grade surgical sepsis.
There were a total number of 1884 patients in the sample group of whom 855 were female (45.4%). The median patient age was 36 years (IQR 23–56). A total of 1489 patients (79%) were deemed to have high-grade sepsis based on an advanced EGS AAST grading, whilst 395 patients (21%) had low-grade sepsis. A total of 71 patients died (3.8%). Of these patients who died, 67 (94.4%) had high-grade sepsis and 4 (5.6%) had low-grade sepsis. The mortality rate in the high-grade sepsis group was 4.5%, whilst the mortality rate in the low-grade sepsis group was 1%. The scores with the greatest accuracy in predicting mortality were qSIRS (AUROC 0.731, 95% CI 0.68–0.78), followed by SIRS (AUROC 0.70, 95% CI 0.65–0.75). The qSOFA and qSOFA L were the least accurate in predicting mortality (AUROC 0.684, 95% CI 0.63–0.74 for both). The addition of lactate had no significant effect on the accuracy of the five scores in predicting mortality. Patients with a qSOFA ≥ 2 have an increased risk of dying (OR 5.8), as do patients with a SIRS score ≥2 (OR 2.7). qSIRS L had the highest sensitivity (69%) in predicting the presence of high-grade surgical sepsis, followed by qSIRS (65.5% sensitivity). qSOFA showed a very low sensitivity of only 4.5% and a high specificity of 99.2%. The addition of lactate to the score marginally improved the sensitivity. Lactate of 2mmol/L or more was also an independent predictor of high-grade sepsis.
The qSIRS score is most accurate in predicting mortality in surgical sepsis. The qSOFA score is inferior to both the SIRS and the qSIRS scores in predicting mortality. The qSIRS score with the addition of lactate to the qSIRS score made it the most sensitive score in predicting high-grade surgical sepsis.