Secondary ischemic events worsen the outcome of patients with severe head injury. Such a secondary ischemic event may be caused by a forced hyperventilation. A consequence of the induced vasoconstriction is the risk of ischemia with an adverse effect on outcome. As a reliable and on-line technique, brain tissue pO2 (p(ti)O2) is used for monitoring regional microcirculation, to detect critical hypoperfusion. On 22 patients with a severe head injury 70 hyperventilation tests were performed from day 0-9 after trauma, calculating TCD-CO2-reactivity (% change of mean flow velocity per mm Hg paCO2 change). Additionally brain p(ti)O2-CO2-reactivity (% change of brain p(ti)O2 per mm Hg paCO2 change) was calculated and introduced. Group A +2 (p(ti)O2 < or = 15 mm Hg, TCD-CO2-reactivity > or = 2.5%, p(ti)O2-CO2-reactivity > 0%) and group B +2 (p(ti)O2 > 15 mm Hg, TCD-CO2-reactivity > or = 2.5%. p(ti)O2-CO2-reactivity > 0%) was formed. P(ti)O2 values in group A+2 decreased to an ischemic level or ischemia aggravated during hyperventilation. In group B+2 no ischemic events occurred. TCD-CO2-reactivity, p(ti)O2-CO2-reactivity and decrease of paCO2 were not significantly different in both groups. 6 out of 22 patients showed, from day 0-9, at least once a risk of (aggravating) ischemia by hyperventilation therapy.