Using noninvasive cerebral oximetry during prehospital triage was a quick way to detect large artery occlusion (LAO), paving the way for the better selection of stroke patients going to comprehensive stroke centers for endovascular therapy, one center reported.
The finding that a relatively large difference in oxygen saturation exists between brain hemispheres (ΔBrSO2) could be an indicator of LAO stroke, according to Alexander Flint, MD, PhD, of Kaiser Permanente in Redwood City, California, and colleagues. There was an 8.3% interhemispheric difference among those who did have an LAO, as opposed to those with versus small artery stroke (0.4%), hemorrhagic stroke (0.4%), and no stroke (0.2%, P<0.001).
Setting a ΔBrSO2 threshold at -3% made it possible to pick out LAOs with 84.2% sensitivity and 70% specificity (area under the curve 0.77, 95% CI 0.67-0.88), they found in their pilot study. When Flint’s group considered the G-FAST clinical score (threshold 3+) in addition to the ΔBrSO2 measure, the new model specificity improved to 90% with no change in sensitivity (AUC 0.88, 95% CI 0.79-0.97), they reported in Stroke.
This is one way to better determine which patients need to go to an endovascular-capable hospital and which are fit for treatment at a primary stroke center, the authors suggested. Endovascular stroke therapy corrected the imbalance in brain oxygen saturation among 16 out of 19 LAO patients.
Near-infrared spectroscopy-based cerebral oximetry was performed using Medtronic’s INVOS monitor in this prospective study of 69 patients. Clinicians at a comprehensive stroke center placed small adhesive sensors on symmetrical forehead locations to detect anterior circulation LAOs.
“The approach tested here is clinically well established, with regulatory approval for many years. The monitor takes only 1 to 2 minutes to obtain measurements and is quite straightforward to deploy (in our experience, <10 minutes of in-service training was required for physicians, nursing staff, or technicians to learn to use the device)," Flint's group wrote.
A pilot study performed in hospitalized patients showed that this experience may necessarily not be applicable to the prehospital triage population. “Next steps to further evaluate this approach will require assessment using unselected suspected stroke subjects in the emergency department and prehospital environments,” according to the investigators.
Another limitation relates to the nature of the device used: hair gets in the way of its sensors, leaving large parts of the head unmeasurable.
Flint and co-authors disclosed no relevant relationships with industry.