Please see Bibliography of References for a list of abstracts, manuscripts and posters.

Sepsis

Accuracy of near infrared spectroscopy in predicting central venous mixed oxygenation in the emergency department

Napoli A, Forcada A, Corl K, Gardiner F. Acad Emerg Med. 2009;16(S1):S8.

Introduction: Invasive measurement of central venous oxygenation (ScvO2) is a key piece of early goal directed therapy (EGDT). Near infrared spectroscopy offers noninvasive, continuous measurement of tissue oxygenation (StO₂).

Objectives: We hypothesized a clinically relevant relationship would exist between StO₂ with ScVO2 in patients undergoing EDGT with an a priori error tolerance of ±5%.

Methods: Continuous measurements of StO₂ (using InSpectra™ at the thenar eminence) and ScVO2 (using the PreSep™ Central Venous Oximetry Catheter) were completed in 30 patients with 226 measurements, none lost to followup. At least 3 concurrent measurements of StO₂ and ScVO2 were made at 15 minute intervals. A random coefficients model with random and fixed intercepts and slopes was used to estimate the relationship between measurements of oxygen saturation taken between the two devices, while accounting for the correlation of measurements taken within the same patient. An unstructured variance-covariance matrix was used and the model was fit using residual estimation of maximum likelihood.

Results: A statistically significant relationship exists between StO₂ and ScVO2 (p<0.001) but poor correlation (r=0.34). The slope of 0.370 (95% CI 0.218-0.523) was statistically significantly different from 1.0, and the intercept of 49.58 (95% CI 37.33-61.84) was statistically significantly different from zero. Only 7% of patients had all measurements fall within the pre-specified acceptable range, while 57% fell completely outside. Eighty seven percent of all measurements had ScVO2.>60%.

Conclusions: This is the largest comparison study of these two technologies in EGDT patients. A significant relationship exists between these technologies but not one that predicts or replaces ScVO2 with StO₂. The limited number of measurements in the lower range of ScVO2 has the potential to create high points of leverage in regression and affect our conclusions so further study in lower ScVO2 populations may clarify this relationship.