CEREBRAL VENOUS OXYGEN-SATURATION STUDIED WITH BILATERAL SAMPLES IN THE INTERNAL JUGULAR VEINS

THE CURRENT LITERATURE reports many measurements (arteriovenous oxygen content difference and cerebral metabolic rate of oxygen, etc.) with samples from the internal jugular veins (IJs), obtained from either si de of the neck, based on the assumption that a reliable sample of mixe d venous blood can be drawn. We compared oxygen saturation in both IJs in 32 patients with head injuries to establish the similarities or di screpancies in the two veins. Both IJs were cannulated with 20-G cathe ters; in five patients, a fiberoptic catheter was used to obtain a con tinuous recording of the hemoglobin saturation. Blood samples were tak en simultaneously from the two IJs and immediately processed; the tota l number of samples processed was 342, with an average of 5.34 paired samples from each patient. The mean and the standard deviation of the differences between the saturation of the two IJs were, respectively, 5.32 and 5.15. Fifteen patients showed differences greater than 15% in hemoglobin saturation; three more patients showed differences greater than 10% at some point during the investigation. Ultimately, only eig ht patients had differences of less than 5%. No relationship was found among the computed tomographic scan data and the pattern of hemoglobi n saturation detected. Therefore, we were not able to identify the sid e more appropriate for monitoring in patients with bilateral, predomin antly monolateral, cortical, or deeply located lesions. The 95% confid ence limits for the percentage of patients with a difference higher th an 15% were between 30 and 64%; the limits for the percentage of patie nts with a difference higher than 10 were between 39 and 73%; ultimate ly, the limits for the percentage of patients with a difference higher than 5 were between 60 and 90%. The proportion of patients with relev ant discrepancies between the two IJs is higher than suspected, and th e reliability of a single item of data obtained from a single IJ is qu estionable.