Intracranial Pressure / Nervous System / Critical Care Manual

The normal brain weighs 1400 g and contains 75 cc of CSF and 75 cc of blood. The cranium can absorb an additional 100-150 cc fluid before ICP begins to rise – this ability to conform increases as patients age (volume may shrink as much as 30%). CSF is produced at 400-500 cc/day.

Unlike most organ systems, whose perfusion pressure is dependent on the difference between MAP and CVP, the brain is often dependent on intracranial pressure (ICP). Why? Because if ICP exceeds CVP, the “driving force” of blood across the intracranial arterioles is MAP – ICP (and not MAP – CVP). The recognition of this led to the concept of Cerebral Perfusion Pressure, defined as MAP – ICP (or CVP if it is higher than ICP). Management of patients with head injuries focuses on optimizing perfusion, i.e., minimizing ICP and maximizing CPP. Complicating this is the fact that excessive cerebral edema can cause herniation, which is fatal independent of the effects of CPP. There is thus considerable debate as to whether ICP or CPP is a more important target (see #ICP vs. CPP for more).

While autoregulation will lead to linear CVR changes with PaCO2 from 20 – 80 mm Hg and CPP from 60 – 160 mm Hg, autoregulation may fail following stroke or traumatic brain injury [J Neurotrauma 9: S333, 1992]. Cerebral edema following stroke or TBI maximizes a few days after the incident. In order to roughly gauge the brain’s compliance, look at the slope of the ICP waveform.

 Outcomes and Intracranial Pressure

Data from the Traumatic Coma Data Bank suggest that ICPs over 20 mm Hg, particularly if sustained, lead to worse outcomes – this has been corroborated by several other large studies [J Neurosurg 69: 15, 1988; J Neurosurg 58: 566, 1983; J Neurosurg 56: 650, 1982], most recently an analysis of 846 TBI patients, which showed that mortality rates were 14% if ICP was < 20 mm Hg by 48 hrs, but 34% if ICP was > 30 mm Hg at 48 hrs [Jiang JY et al. J Neurotrauma 19: 869, 2002]. That said, the Brain Trauma Foundation’s recommendation to initiate intracranial pressure lowering at ICP thresholds of 20 mm Hg is only a Level II recommendation [Bratton SL et al. J Neurotrauma 24 (S1): S55-S58, 2007]

Recent studies have shown that brain tissue oxygenation and metabolism are adversely affected by CPP less than 60-70 mm Hg [Acta Neuochir Suppl (Wien) 71: 153, 1998; J Trauma 42: S32, 1997; J Neurosurg 89: 971, 1998] – where in this range CPP needs to stay is controversial, with a consensus that the range should be 60-70 [J Neurotrauma 13: 639, 1996]. The Brain Trauma Foundation Guidelines provide no Level I recommendations on the subject, provide a Level II recommendation that attempts to exceed a CPP of 70 mm Hg are counterproductive, and provide a Level III recommendation that a CPP of < 50 mm Hg should be avoided [Bratton SL et al. J Neurotrauma 24 (S1): S59-S64, 2007]

Andrews recommends continuous monitoring of temperature (keep < 38C) in addition to hemodynamic variables.

Treatment of Elevated Intracranial Pressure

The use of sedatives to lower ICP is controversial – in the absence of agitation or anxiety there is no clear evidence that paralysis or sedation are beneficial. In fact, data from the Traumatic Coma Data Bank suggest that general use of sedation did not improve outcomes but increased complications and lengthened ICU stay [Crit Care Med 22: 1471, 1994]. On the other hand, anxiety, agitation, or spontaneous posturing can raise ICP and should be treated with morphine at 2-5 mg/kg/hr and vecuronium at 10 mg/hr.

CSF Drainage

CSF drainage can be highly effective, even if only a small amount is removed. Continuous drainage is NOT recommended as it precludes ICP monitoring and has been shown to increase the risk of catheter malfunction (secondary to closed ventricular walls).


Mannitol boluses at 0.25-1g/kg are preferred over continuous infusion as the infusion increases uptake into brain tissue, reversing the osmotic gradient and potentially causing harm. Mannitol works by both improved osmotic gradient and rheology – studies have shown that mannitol increases CBF by as much as 20% and decreases CBF after brain injury [J Neurosurg 61: 700, 1984], although it was recently shown not to affect brain tissue oxygenation [J Trauma 62: 292, 2007]. Note that there is some evidence that Osm > 320 will impair renal function [NEJM 297: 1449, 1977], and most people will hold mannitol of [Na+] > 155-160, however the data supporting either of these is not strong.

A series of 8 patients who went into ARF s/p mannitol showed failure within 3.5 +/- 1.1 days after mannitol doses of 189 +/- 64 g daily (626 +/- 270 g total). Peak osmolal gap was 74 +/- 39 mOsm/kg water. In patients with normal baseline renal function ARF developed after receiving total mannitol doses of 1171 +/- 376 g. The peak osmolal gap was 107 +/- 17 (ie measured serum osmolality 376). In those with underlying renal compromise, renal function worsened after a total mannitol dose of 295 +/- 143 g [Medicine (Baltimore) 69: 153, 1990]. Limited data suggest that prolonged Osm > 320 mOsm/L are associated with a higher mortality [Acta Neurochir (Wien) 116: 33, 1992].

Much of this is refuted by a recent retrospective study of 98 patients on mannitol at WUSTL, the multivariate analysis of which showed that APACHE II and history of CHF were the only predictive factors leading to mannitol-induced renal failure. Osmolality gap and mannitol dose had no correlation. Furthermore, all cases of MI-RF reversed [J Neurosurgery 103: 244, 2005] Loop diuretics are helpful as well but are also recommended only when Osm < 320 and [Na+] < 155


Hyperventilation is known to lower ICP however CBF drops 3-4% for every 1 mm Hg decrease in PCO2 [Arch Neurol 23: 394, 1970] – this is dangerous as CBF may drop by as much as 50% following TBI [J Neurosurg 74: 407, 1991; J Neurosurg 75: 685, 1991]. Hyperventilation is highly controversial [Chest 127: 1812, 2005], with the 2007 Cochrane Database Review concluding that there is inadequate data to assess whether benefit or harm exists [Roberts, I; Schierhout, G: Cochrane Review 2007]. The Brain Trauma Foundation recommends against chronic hyperventilation [J Neurotrauma 17: 537, 2000] – Andrews recommends 35 mm Hg.


Barbiturates have been studied by in several prospective, randomized clinical trials with none of them showing a clear benefit [Can J Neurol Sci 11: 434, 1984; J Neurosurg 62: 383, 1985], however one suggested a benefit in patients in whom barbiturates lowered ICP [J Neurosurg 69: 15, 1998] – a 2000 Cochrane Database Review, however, concluded that “there is no evidence that barbiturate therapy in patients with acute severe head injury improves outcome. Barbiturate therapy results in a fall in blood pressure in 1 in 4 treated patients. The hypotensive effect of barbiturate therapy will offset any ICP lowering effect on cerebral perfusion pressure” [Cochrane Database Syst Rev. 2000;(2):CD000033]. If patients are tried on barbiturates, wean them as soon as possible to avoid myocardial complications as well as pneumonia, for which these patients are at a high risk.


There is some data to suggest that moderate hypothermia (32-33C) can reduce ICP [Neurol Res 20S1: S52, 1998; Neurosurgery 21: 177, 1987; J Neurosurg 85: 533, 1996; J Neurosurg 79: 363, 1995]. One trial suggested a benefit in terms of early outcomes, but this dissipated by 12 months [NEJM 336: 540, 1997]

Key Points: Intracranial Pressure

  • Multiple studies confirm that sustained ICP > 20 mm Hg worsens outcome
  • Brain tissue oxygenation and metabolism are adversely affected by CPP < 60-70 mm Hg
  • Elevate head of bed and prevent venous outflow obstruction
  • Sedation should only be used in the agitated or anxious patient [Crit Care Med 22: 1471, 1994]
  • Mannitol 0.25-1.0 g/kg in boluses lowers ICP and improves CBF, however the Osm < 320 and Na < 155 requirement is not based on solid data. Newer data suggest that mannitol induced renal failure is correlated with APACHE-II score and CHF, and has nothing to do with osmole gap or mannitol dose [J Neurosurgery 103: 244, 2005]
  • Hyperventilation is controversial, most recommend no lower than 35 mm Hg
  • There is no data that barbiturates improve outcome [Cochrane Database 000033]
  • There is no data that hypothermia improves outcome


A retrospective look at 427 in the NMDA antagonist Selfotel trial found that the most powerful predictor of neurological worsening was ICP  20 mm Hg either initially or during neurological deterioration. There was no correlation with the CPP as long as the CPP was > 60 mm Hg [J Neurosurg 92: 1, 2000]

The only class II evidence on this subject is a randomized clinical trial of 189 comatose adults s/p severe head injury. All patients maintained ICP < 20 mm Hg, the difference was in how this was achieved – in the CBF-targeted group (MAP > 90 mm Hg, CPP > 70 mm Hg, and PaCO2 ~ 35 mm Hg) MAPs were kept > 90 mm Hg and hyperventilation was not used. In the ICP-targeted group (MAP > 70 mm Hg, and CPP > 50 mm Hg), hyperventilation to a PaCO2 of 25-30 mm Hg used and MAPs were kept > 70 mm Hg. The CBF protocol reduced jugular desaturation from 50.6% to 30% (p = 0.006). When the frequency of jugular desaturation was adjusted for all confounding factors that were significant, the risk of cerebral ischemia was 2.4-fold greater with the ICP-targeted protocol. Despite the reduction in secondary ischemic insults, there was no difference in neurologic outcome, probably because 1) jugular desaturation is easily treated and 2) the CBF group suffered a five-fold increase in the frequency of ARDS [Robertson CS et al. Crit Care Med 27: 2086, 1999]

Brain Trauma Foundation Guidelines

Normal adult CPP is 50 mm Hg. The Brain Trauma Foundation Guidelines claim the following – there is no class I evidence to make any recommendations. Class II evidence suggests avoiding CPP > 70 mm Hg (risk of pulmonary edema is too high). Class III evidence suggests that CPP < 50 mm Hg should be avoided [J Neurotrauma 24: S59, 2007]

Key Points: CPP vs ICP in Traumatic Brain Injury

  • As long as CPP > 60 mm Hg, ICP control appears to be more important than further increases in CPP (Class III data, [J Neurosurg 92: 1, 2000])
  • In order to keep ICP < 20 mm Hg, MAP 90/CPP 70 vs. MAP 70/CPP 50 are no different in terms of neurologic outcome [Crit Care Med 27: 2086, 1999]