Fifty-five head injured patients (GCS < 8) were studied at an average of 7.5 +/- 3.4 days on the ICU to check quality of hemodynamic monitoring and the consequences for therapy. Multimodal neuromonitoring included intracranial pressure (ICP), mean arterial pressure (MAP), cerebral perfusion pressure (CPP), endtidal CO2 (EtCO2) as well as brain tissue--pO2 (p(ti)O2), regional oxygen (rSO2) and jugular venous oxygen saturation (SjO2). Regional p(ti)O2 as well as global SjO2 were sensitive technologies to detect hemodynamic changes. However analyzing reliability and good data quality regional p(ti)O2 (up to 95%) was superior to jugular bulb oximetry (up to 50%). Longterm-measurements of rSO2 using near infrared spectroscopy reached, if possible, a restricted reliability (good data quality up to 70%) and sensitivity in comparison to p(ti)O2. Especially p(ti)O2 enabled detection of critical p(ti)O2 (< 15 mm Hg) in up to 50% frequency during the first days after trauma and a second peak after day 6 to 8 according to evidence of CPP insults. Knowledge of baseline p(ti)O2 and CO2-reactivity allowed minimizing risk of ischemia by induced hyperventilation and improvement on cerebral microcirculation after mannitol administration could be individually recognized.