Measurement of Cerebrovascular Reactivity as Blood Oxygen Level-Dependent Magnetic Resonance Imaging Signal Response to a Hypercapnic Stimulus in Mechanically Ventilated Patients

Background Impaired cerebrovascular reactivity (CVR) is an important prognostic marker of stroke. Most measures of CVR lack (1) a reproducible vasoactive stimulus and (2) a high time and spatial resolution measure of cerebral blood flow (CBF), particularly for mechanically ventilated patients. The aim of our study was to investigate the feasibility of measuring CVR using sequential gas delivery circuit and gas blender for precise targeting of end-tidal PCO 2 (P et CO 2 ), and blood oxygen level-dependent magnetic resonance imaging (BOLD-MRI) signal as a surrogate of CBF, in mechanically ventilated patients. Methods Four patients with known moyamoya disease requiring preoperative CVR measurements under general anesthesia were studied. All patients had standard anesthesia induction and maintenance with intravenous propofol and rocuronium. Patients were intubated and manually ventilated with a self-inflating bag connected to a sequential breathing circuit. A computer-controlled gas blender supplied the gas mixture in proportions to attain target P et CO 2 . BOLD-MRI was performed at 3.0 Tesla magnet. Changes in signal per change in P et CO 2 were calculated, and their magnitude color-coded and mapped onto the anatomic scan to form CVR maps. Results CVR studies were successfully performed on all patients, and the CVR values were lower in both gray and white matter bilaterally when compared with healthy volunteers. In addition, CVR maps in 3 patients showed intracerebral steal phenomenon in spite of having had cerebral revascularization procedures, indicating that they are still at risk of cerebral ischemia. Conclusions BOLD-MRI CVR studies are feasible in mechanically ventilated patients anesthetized with propofol.