[Near-infrared spectroscopy in sepsis therapy : predictor of a low central venous oxygen saturation].

BACKGROUND: Early goal-directed hemodynamic optimization has become a cornerstone of sepsis therapy. One major defined goal is to achieve adequate central venous oxygen saturation (SO(2)). This study aimed to investigate the correlation between central venous SO(2) and frontal cerebral near-infrared spectroscopy (NIRS) measurement in patients with severe sepsis and septic shock. The NIRS method provides non-invasive measurement of regional oxygen saturation (rSO(2)) in tissues approximately 2 cm below the optical NIRS sensors which depends on arterial, capillary and venous blood. Thus this system gives site-specific real-time data about the balance of oxygen supply and demand.

METHODS: This was a secondary analysis from a prospective study of surgical intensive care (ICU) patients in the early phase of severe sepsis or septic shock. Bilateral cerebral rSO(2), central venous SO(2), arterial oxygen saturation (S(a)O(2)) and other surrogate parameters of oxygen supply, such as hemoglobin, partial pressure of oxygen and oxygen content in arterial blood were recorded.

RESULTS: A total of 16 ICU patients (4 women, median age 65.5 years) were included in the study. As sepsis focus an intra-abdominal infection was detected in 62.5 % of patients, severe pneumonia was determined in 31.3 % and skin and soft tissue infections were recognized in 12.5 %. At study inclusion 50 % of patients had septic shock, the median sequential organ failure assessment (SOFA) score was 10.2 (interquartile range 5.25-8.75) and the median acute physiology and chronic health evaluation II (APACHE II) score was 26 (range 23.25-29.75). Mortality at day 28 was 37.5 %. Minimum rSO(2) (median 58) and right-sided rSO(2) (median 58) values showed a significant correlation in the analysis of receiver operating characteristics (area under the curve 0.844, p= 0.045). A central venous SO(2)< 70 % was indicated by rSO(2)< 56.5 with sensitivity and specificity of 75 % and 100 %, respectively.

CONCLUSIONS: Cerebral NIRS could provide a fast and easily available side effect-free monitoring that could be used in addition to established procedures for goal-directed treatment in the early phase of sepsis. Further studies should be made in a larger population to verify the correlation found and to investigate the impact of NIRS-directed resuscitation treatment in early sepsis.