Near-infrared spectroscopy to predict organ failure and outcome in sepsis: the Assessing Risk in Sepsis using a Tissue Oxygen Saturation (ARISTOS) study

OBJECTIVES: Sepsis is acute organ dysfunction in the setting of infection. An accurate diagnosis is important to guide treatment and disposition. Tissue oxygen saturation (StO2) can be estimated noninvasively by near-infrared spectroscopy (NIRS), and may be an indicator of microcirculatory dysfunction in early sepsis. We aimed to determine the utility of StO2 for sepsis recognition and outcome prediction among patients presenting to the emergency department (ED) with infection. PATIENTS AND METHODS: A multicentre, prospective, observational cohort study recruited patients who were being admitted to hospital with infection. StO2 was measured in the ED using a handheld NIRS device, Inspectra 300. Outcomes were sepsis, defined as an increase in sequential organ failure assessment score of at least 2 points within 72 h, and composite in-hospital mortality/ICU admission at least 3 days. RESULTS: A cohort of 323 participants, median age 64 (interquartile range: 47-77) years, was recruited at three Australian hospitals. 143 (44%) fulfilled the criteria for sepsis and 22 (7%) died within 30 days. The mean ± SD StO2 was 74 ± 8% in sepsis and 78 ± 7% in nonsepsis (P < 0.0001). StO2 correlated with the peak sequential organ failure assessment score (Spearman's ρ -0.27, P  <  0.0001). Area under the receiver operating characteristic curve was 0.66 (95% confidence interval: 0.60-0.72) for sepsis and 0.66 (0.58-0.75) for the composite outcome. StO2 less than 75% had an odds ratio of 2.67 (1.45-4.94; P = 0.002), for the composite outcome compared with StO2 at least 75%. CONCLUSION: NIRS-derived StO2 correlates with organ failure and is associated with outcome in sepsis. However, its ability to differentiate sepsis among ED patients with infection is limited. NIRS cannot be recommended for this purpose.