OBJECTIVE Traditionally, intracranial pressure (ICP) monitoring has been utilized in all patients with severe head injury (Glasgow coma score of 3-8). Ventriculostomy placement, however, does carry a 4 to 10 percent complication rate consisting mostly of hematoma and infection. The authors propose that a subgroup of patients presenting with severe head trauma and diffuse axonal injury without associated mass lesion, do not need ICP monitoring. Additionally, the monitoring data from ICP, MAP, and CPP for a comparison severe head injury group, and subgroups of DAI would be presented. METHODS Thirty-six patients sustaining blunt head trauma and fitting our strict clinical and radiographic diagnosis of DAI were enrolled in our study. Inclusion criteria were severe head injury patients who did not regain consciousness after the initial impact, and whose CT scan demonstrated characteristic punctate hemorrhages of < 10 mm diameter at the greywhite junction, basal ganglia, corpus callosum, upper brainstem, or a combination of the above. Patients with significant mass lesions and documented anoxia were excluded. Their intracranial pressure (ICP) and cerebral perfusion pressure (CPP) were compared to a control group of 36 consecutive patients with severe non-penetrating non-operative head injury, using the Analysis for Variance method. RESULTS Eighteen (50.0%), six (16.7%), and twelve (33.3%) patients had types I, II, and III DAI, respectively. The admission Glasgow Coma Score (GCS) was higher for types I and II than for type III DAI. ICP was monitored from 23 to 165 hours, with a mean ICP for 36 patients of 11.70 mmHg (SEM = 0.75) and a range from 4.3 to 17.3 mmHg. Of all ICP recordings, of which 89.7%, (2421/2698) were < or = 20 mmHg. Average mean arterial pressure (MAP) was 96.08 mmHg (SEM = 1.69), and 94.6% (2038/2154) of all MAP readings were greater than 80 mmHg. Average cerebral perfusion pressure (CPP) was 85.16 mmHg (SEM = 1.68), and 90.1% (1941/2154) of all CPP readings were greater than 70 mmHg. This is compared to the control group mean ICP, MAP, and CPP of 16.84 mmHg (p = 0.000021), 92.80 mmHg (p = 0.18), and 76.49 mmHg (p = 0.0012). No treatment for sustained elevated ICP > 20 mmHg was needed for DAI patients except in two; one with extensive intraventricular and subarachnoid hemorrhage who developed communicating hydrocephalus, and another with ventriculitis requiring intrathecal and intravenous antibiotic treatments. Two complications, one from a catheter tract hematoma, and another with Staph epidermidis ventriculitis, were encountered. All patients, except type III DAI, generally demonstrated marked clinical improvement with time. The outcome, as measured by Glasgow Coma Score (GCS) and Glasgow Outcome Score (GOS) was similarly better with types I and II than type III DAI. CONCLUSIONS The authors conclude that ICP elevation in DAI patients without associated mass lesions is not as prevalent as other severe head injured patients, therefore ICP monitoring may not be as critical. The presence of an ICP monitoring device may contribute to increased morbidity. Of key importance, however, is an accurate clinical history and interpretation of the CT scan.