Intubation, which requires sedation and myorelaxants, may lead to inaccurate neurological evaluation of severely head-injured patients. Aims of this study were to describe the early clinical evolution of traumatic brain injured (TBI) patients admitted to intensive care unit (ICU), to identify cases of over-estimated neurological severity, and to quantify the risk factors for this over-estimation. A total of 753 TBI patients consecutively admitted to ICU of three academic neurosurgical hospitals (NSH) were assessed. Cases whose severity was potentially over-estimated were identified by four criteria and indicated as "mistakenly severe" (MS): (1) no surgical intracranial masses; (2) could not follow commands at neurological assessment; (3) were dismissed from the ICU in < or =3 days to a regular ward; and (4) had regained the ability to obey commands. A total of 675 patients were intubated and/or sedated-paralyzed at the post-stabilization evaluation. In all, 304 patients had surgically treated intracranial masses. Among the 449 non-surgical cases, 58 patients fulfilling the criteria for MS were identified. The main features distinguishing MS from truly severe cases were younger age, higher Glasgow Coma Scale (GCS) score at all time points, Marshall classification of Computerized Tomographic (CT) scan mostly Diffuse Injury I and II, fewer pupillary abnormalities, and a lower frequency of hypoxia, hypotension, and extra-cranial injuries. In a certain proportion of non-surgical TBI patients, mostly intubated and sedated, neurological examination is difficult and severity can be over-estimated. Risk factors for this inaccurate evaluation can be identified, and clinical decisions should be based on further examination.