Objective To investigate the efficacy of early, routine use of neuromuscular blocking agents for intracranial pressure management in patients with severe head injury. Design Retrospective review of data from the Traumatic Coma Data Bank. The Traumatic Coma Data Bank was a collaborative project of the National Institute of Neurological Disorders and Stroke that involved four Level I trauma centers. The Traumatic Coma Data Bank prospectively collected data on 1,030 severe head-injured patients (Glasgow Coma Score of ≥8) between 1984 and 1987. Setting Four Level I trauma centers coordinated by the National Institute of Neurological Disorders and Stroke. Patients A total of 514 Traumatic Coma Data Bank patients who met study inclusion criteria were divided into two groups: group 1 consisted of 239 patients who were pharmacologically paralyzed starting within the first intensive care unit (ICU) shift or ≤6 hrs into the second shift and lasting for at least 12 hrs. Group 2 contained the remaining 275 patients who did not fulfill these criteria. Measurements and Main Results We analyzed the acute care course of these patients from the time of injury through admission, resuscitation, and ICU stay, as well as their rehabilitation follow-up results for ≤1 yr. These two groups did not differ significantly regarding admission intracranial computed tomography diagnosis, admission Glasgow motor score, percentage of monitored time that intracranial pressure was ≥20 mm Hg, or frequency of initial intracranial pressure ≥20 mm Hg. In group 2, more patients were aged >40 yrs (24% vs. 15% in group 1) or had admission hypotension (33% in group 2 vs. 25% in group 1). Multivariate regression analysis showed that ICU stay was significantly longer in group 1 (mean 7.76 days) than group 2 (mean 4.84 days; p < .001). The occurrence of pneumonia was significantly higher in group 1 (29% vs. 15%, p < .001). There was a tendency toward a higher occurrence of sepsis in group 1, but this difference did not reach statistical significance (11% vs. 7%, p = .08). The final Glasgow Outcome Scale scores for the two groups were not significantly different in those survivors with good and moderate outcomes. Although there were more deaths in group 2 (39% vs. 24%, p < .001), there were more vegetative or severely disabled survivors in group 1 (8% vs. 4% and 21% vs. 13%, respectively). Conclusions Our findings suggest that early, routine, long-term use of neuromuscular blocking agents in patients with severe head injuries to manage intracranial pressure does not improve overall outcome and may actually be detrimental because of the prolongation of their ICU stay and the increased frequency of extracranial complications associated with pharmacologic paralysis. We suggest that routine early management of the head-injured patient in the ICU should be accomplished using sedation alone and that neuromuscular blockade should be generally reserved for patients with intracranial hypertension who require escalation of treatment intensity. (Crit Care Med 1994; 22:1471–1476)