Performance of Quick Sepsis-Related Organ Failure Assessment (qSOFA) in Predicting Adverse Outcomes in Acute Leukemia Patients Receiving Induction Chemotherapy

Background: The Quick Sepsis-related Organ Failure Assessment (qSOFA) was recently introduced with the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) as a way to quickly identify patients at risk of suffering adverse outcomes from infection. qSOFA is a bedside scoring tool that assigns one point each for altered mentation, respiratory rate of 22/min or greater, and systolic blood pressure of 100 mmHg or less. Scores of 2 or greater are associated with increased rates of ICU admission and in-hospital mortality. Although it has been validated in general emergency department and hospital ward patients, its performance in cancer patients remains unknown. The aim of this study was to evaluate the performance of qSOFA in predicting adverse outcomes in patients with acute leukemia receiving induction chemotherapy.

Methods: We reviewed the electronic health records of all patients with acute leukemia who were admitted for induction chemotherapy between October of 2011 and December of 2015 at our institution. Clinical variables included demographic information, type of leukemia, type of chemotherapy, vital signs, and assessment of mental status by the Richmond Agitation-Sedation Scale. The primary endpoints were shock and in-hospital mortality. The qSOFA score was calculated for each patient for each day of their hospitalization using the worst physiologic parameters for each day. The maximum score for each patient from 48 hours prior to the development of clinical infection to 24 hours post development was then included in subsequent analysis. Clinical infection was defined as a positive culture from any source, and shock was defined as a vasopressor requirement to maintain a mean arterial pressure of ≥ 65 mmHg.

Results: 572 patients with acute leukemia were admitted for induction chemotherapy between October of 2011 and December of 2015. Three hundred ninety patients (68%) had acute myeloid leukemia, 114 (20%) had acute lymphoblastic leukemia, 55 (10%) had acute promyelocytic leukemia, and 13 (2%) had mixed phenotypes. Among those patients admitted for induction chemotherapy, 212 (42%) developed clinical infection during their hospitalization. Further analysis revealed that 32 (15%) clinically infected patients developed shock and 14 (7%) died in the hospital. Using the existing cut-off score of 2, sensitivity was 88% (95% Confidence Interval [CI]: [71%, 97%]) and specificity was 64% (95% CI: [57%, 71%]) for detecting shock. For mortality, sensitivity was 79% (95% CI: [49%, 95%]) and specificity was 59% (95% CI: [52%, 66%]). Considering scores continuously, the area under the receiver operating characteristic curve for shock was 0.84 (95% CI: [0.77, 0.91]) and was 0.76 (95% CI: [0.63, 0.88]) for mortality.

Conclusion: qSOFA is a bedside scoring tool that may help clinicians predict adverse outcomes in acute leukemia patients receiving induction chemotherapy who are suspected of having infection. However, it is limited by relatively low specificity in this subset of patients, which may be due the confounding effect that intensive chemotherapy has on vital signs and mental status. Future studies should focus on developing scoring systems with less false positives in this population.