Object: The aim of this study was to investigate the effects of moderate hyperventilation on intracranial pressure (ICP), jugular venous oxygen saturation ([SjvO2], an index of global cerebral perfusion), and brain tissue PO2 (an index of local cerebral perfusion).
Methods: Ninety-four tests consisting of 20-minute periods of moderate hyperventilation (27-32 mm Hg) were performed on different days in 36 patients with severe traumatic brain injury (Glasgow Coma Scale score < or = 8). Moderate hyperventilation resulted in a significant reduction in average ICP, but in seven tests performed in five patients it was ineffective. The response of SjvO2 and brain tissue PO2 to CO2 changes was widely variable and unpredictable. After 20 minutes of moderate hyperventilation in most tests (79.8%), both SjvO2 and brain tissue PO2 values remained above the lower limits of normality (50% and 10 mm Hg, respectively). In contrast, in 15 tests performed in six patients (16.6% of the studied population) brain tissue PO2 decreased below 10 mm Hg although the corresponding SjvO2 values were greater than 50%. The reduction of brain tissue PO2 below 10 mm Hg was favored by the low prehyperventilation values (10 tests), higher CO2 reactivity, and, possibly, by lower prehyperventilation values of cerebral perfusion pressure. In five of those 15 tests, the prehyperventilation values of SjvO2 were greater than 70%, a condition of relative hyperemia. The SjvO2 decreased below 50% in four tests; the corresponding brain tissue PO2 values were less than 10 mm Hg in three of those tests, whereas in the fourth, the jugular venous O2 desaturation was not detected by brain tissue PO2. The analysis of the simultaneous relative changes (prehyperventilation - posthyperventilation) of SjvO2 and brain tissue PO2 showed that in most tests (75.5%) there was a reduction of both SjvO2 and brain tissue PO2. In two tests moderate hyperventilation resulted in an increase of both SjvO2 and brain tissue PO2. In the remaining 17 tests a redistribution of the cerebral blood flow was observed, leading to changes in SjvO2 and brain tissue PO2 in opposite directions. CCONCLUSIONS. Hyperventilation, even if moderate, can frequently result in harmful local reductions of cerebral perfusion that cannot be detected by assessing SjvO2. Therefore, hyperventilation should be used with caution and should not be considered safe. This study confirms that SjvO2 and brain tissue PO2 are two parameters that provide complementary information on brain oxygenation that is useful to reduce the risk of secondary damage. Changes in SjvO2 and brain tissue PO2 in opposite directions indicate that data obtained from brain tissue PO2 monitoring cannot be extrapolated to evaluate the global cerebral perfusion.