Article Text

Download PDFPDF

Effect of mannitol and hypertonic saline on cerebral oxygenation in patients with severe traumatic brain injury and refractory intracranial hypertension
  1. M Oddo1,
  2. J M Levine1,2,3,
  3. S Frangos1,
  4. E Carrera4,
  5. E Maloney-Wilensky1,
  6. J L Pascual5,
  7. W A Kofke1,3,
  8. S A Mayer4,
  9. P D LeRoux1
  1. 1
    Department of Neurosurgery, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA
  2. 2
    Department of Neurology, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA
  3. 3
    Department of Anesthesiology and Critical Care, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA
  4. 4
    Department of Neurology, Critical Care Division, Columbia University Medical Center, New York, New York, USA
  5. 5
    Department of Surgery, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA
  1. P D LeRoux, Department of Neurosurgery, Clinical Research Division, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA; Peter.LeRoux{at}


Background: The impact of osmotic therapies on brain oxygen has not been extensively studied in humans. We examined the effects on brain tissue oxygen tension (PbtO2) of mannitol and hypertonic saline (HTS) in patients with severe traumatic brain injury (TBI) and refractory intracranial hypertension.

Methods: 12 consecutive patients with severe TBI who underwent intracranial pressure (ICP) and PbtO2 monitoring were studied. Patients were treated with mannitol (25%, 0.75 g/kg) for episodes of elevated ICP (>20 mm Hg) or HTS (7.5%, 250 ml) if ICP was not controlled with mannitol. PbtO2, ICP, mean arterial pressure, cerebral perfusion pressure (CPP), central venous pressure and cardiac output were monitored continuously.

Results: 42 episodes of intracranial hypertension, treated with mannitol (n = 28 boluses) or HTS (n = 14 boluses), were analysed. HTS treatment was associated with an increase in PbtO2 (from baseline 28.3 (13.8) mm Hg to 34.9 (18.2) mm Hg at 30 min, 37.0 (17.6) mm Hg at 60 min and 41.4 (17.7) mm Hg at 120 min; all p<0.01) while mannitol did not affect PbtO2 (baseline 30.4 (11.4) vs 28.7 (13.5) vs 28.4 (10.6) vs 27.5 (9.9) mm Hg; all p>0.1). Compared with mannitol, HTS was associated with lower ICP and higher CPP and cardiac output.

Conclusions: In patients with severe TBI and elevated ICP refractory to previous mannitol treatment, 7.5% hypertonic saline administered as second tier therapy is associated with a significant increase in brain oxygenation, and improved cerebral and systemic haemodynamics.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Intracranial hypertension is common after severe traumatic brain injury (TBI) and may adversely affect outcome.1 2 Control of intracranial pressure (ICP) is therefore a mainstay of treatment after severe TBI. Osmotherapy is frequently used to control ICP. Although no firm recommendations exist, mannitol is more frequently used as a first tier therapy for elevated ICP while hypertonic saline (HTS) is given as a secondline therapy in patients unresponsive to mannitol therapy.3 The comparative effects of these two agents on cerebral physiology, rather than ICP alone, after severe TBI are only beginning to be elucidated. While some authors report that mannitol and HTS have a similar effect, at least when given in equimolar doses,4 others have demonstrated that HTS may be more effective than mannitol in reducing elevated ICP in patients with severe TBI.5 6 In addition to its potent osmotic effect, HTS has beneficial effects on vascular tone,7 and in animal models of TBI, HTS improves systemic haemodynamics,8 9 cerebral blood flow (CBF)911 and may enhance cerebral microcirculation by reducing the adhesion of polymorphonuclear cells12 13 and by stimulating local release of nitric oxide.14

Although ICP and cerebral perfusion pressure (CPP) traditionally are the major targets of TBI treatment, the interstitial partial pressure of oxygen in brain tissue (PbtO2) is emerging as an additional complementary therapeutic target. Specialised sensors placed directly into brain parenchyma allow for continuous bedside assessment of PbtO2 and for the quantification of secondary hypoxic events that occur after the initial brain insult.15 Observational clinical studies demonstrate a relationship between reduced PbtO2 and poor outcome16 17 and suggest that PbtO2 targeted therapy may improve clinical outcomes.18 It is therefore important to understand how various treatments options impact on PbtO2.

There has been limited study on the effect of different osmotic therapies on PbtO2 in brain injured patients and the results vary. For example, in patients with TBI with elevated ICP, mannitol does not consistently improve PbtO2.12 19 In contrast, in patients with intracranial hypertension after subarachnoid haemorrhage, HTS may increase CBF20 21 and PbtO2.22 However, little is known about the impact of HTS on PbtO2 in patients with severe TBI and intracranial hypertension, and in particular those with elevated ICP refractory to mannitol. In this study, we examined how mannitol and HTS, used to treat recurrent episodes of elevated ICP, influenced brain oxygen in patients with severe TBI.


Patient population

Consecutive patients with severe TBI admitted to the Hospital of the University of Pennsylvania, a level I trauma centre, and who underwent PbtO2 monitoring in the neurointensive care unit were retrospectively identified from a prospective observational database (the Brain Oxygen Monitoring Outcome study) over a 2 year period (2005–2006). Severe TBI was defined by (1) a history of trauma, (2) a post-resuscitation admission Glasgow Coma Scale score ⩽8 and (3) clinical and radiographic exclusion of alternate causes of coma. Patients who received osmotherapy with both mannitol and HTS to treat refractory intracranial hypertension (defined as the occurrence of recurrent episodes of ICP >20 mm Hg for more than 10 min despite initial medical management) were included in the analysis. Patients who had bilateral fixed and dilated pupils at admission were excluded from the study. Our Institutional Review Board approved the study.

Intracranial and systemic monitors

ICP, brain temperature and PbtO2 were continuously monitored using commercially available products (Licox, Integra Neuroscience, Plainsboro, New Jersey, USA). Intracranial monitors were inserted at the bedside in the neurointensive care unit through a burr hole into the frontal lobe and secured with a triple lumen bolt. The monitors were placed into white matter that appeared normal on head CT and on the side of maximal pathology. When there was no asymmetry in brain pathology on CT, the probes were placed in the right frontal region. If the patient had undergone a craniotomy, the probes were placed on the same side as the injury if the craniotomy flap permitted. Follow-up head CT scans were performed in all patients within 24 h of admission to confirm correct placement of the various monitors (eg, not in a contusion or infarct). Probe function and stability was confirmed by an appropriate PbtO2 increase following an oxygen challenge (inspired O2 fraction (FiO2) 1.0 for 5 min). PbtO2 measurements were corrected for brain temperature fluctuations. To allow for probe equilibration, data from the first 6 h after PbtO2 monitor insertion were discarded.

Each patient had an indwelling arterial (radial artery) catheter. Heart rate, blood pressure (through the arterial line) and arterial oxygen saturation (SaO2) were recorded continuously in all patients. CPP was calculated from the measured parameters (CPP = MAP–ICP (corrected for ICP and arterial catheter position)). Patients had a pulmonary artery catheter, and central venous pressure and cardiac output (thermodilution) were recorded. As part of routine care, FiO2, SaO2, respiratory rate and ventilator settings (eg, ventilator mode, tidal volume, minute ventilation and positive end expiratory pressure) were recorded in the ICU flowsheet every 15 min. Arterial blood gas analysis was usually performed at 08:00 and 20:00 each day while the patient was ventilated and if there was any significant cardiopulmonary change or at the discretion of the neuro-intensivist. Arterial samples were analysed for haemoglobin, arterial oxygen (PaO2) and carbon dioxide (PaCO2) tension and pH.

General patient management

All patients were managed in the neurointensive care unit according to a local algorithm based on the Brain Trauma Foundation TBI guidelines.3 This included early evacuation of space occupying mass lesions in the operating room. Each patient was fully resuscitated according to Advanced Trauma Life Support guidelines from the American College of Surgeons, intubated and mechanically ventilated with the head of the bed initially elevated ∼20–30°. FiO2 and minute ventilation were adjusted to maintain SaO2 >93%, PaO2 between 90 and 100 mm Hg and PaCO2 between 34 and 38 mm Hg. Volume resuscitation was achieved with 0.9% normal saline and albumin for a target central venous pressure (CVP) of 6–10 cm H2O. After adequate fluid resuscitation, CPP was kept above 60 mmHg, using vasopressors if required. Vasopressors (mainly phenylephrine) were only used to ensure adequate CPP, and patients who received vasopressors for cardio-circulatory failure were excluded from the present study.

Management of elevated ICP

A standard stair step approach was used to treat intracranial hypertension. Therapeutic targets were adjusted to maintain ICP <20 mm Hg and CPP >60 mm Hg. Initial management consisted of head of bed elevation, sedation (lorazepam), analgesia (fentanyl), muscle paralysis (vecuronium) and intermittent cerebrospinal fluid drainage using an external ventricular drain. Optimised moderate hyperventilation (PaCO2 30–35 mm Hg) was used selectively to control ICP provided PbtO2 did not decrease during this intervention.


If ICP remained >20 mm Hg for more than 10 min despite the initial management, osmotherapy was started, provided that serum osmolarity was <320 mosmol and serum sodium <155 mmol/l. In all 12 patients, mannitol (25%, 0.75 g/kg, 412 mosmol/dose, infused over 20 min) was administered as firstline treatment of intracranial hypertension. HTS (7.5% solution, 250 ml, 641 mosmol/dose, infused over 30 min) was used as a secondline therapy to control ICP. The decision to infuse HTS was made at the discretion of the treating neuro-intensivist and based on each patient’s overall therapeutic intensity. However, according to our management protocol, HTS could only be used if a patient had already received mannitol for a previous episode of increased ICP or had a MAP ⩽90 mm Hg. HTS was contraindicated if CVP was >15 mm Hg, or the patient had chronic hyponatraemia, heart failure or diabetes insipidus.

Data collection and analysis

All physiological variables (ICP, brain temperature, MAP, CPP, CVP, SaO2 and PbtO2) were recorded continuously using a bedside monitor (Component Monitoring System M1046-9090C; Hewlett Packard, Andover, Massachusetts, USA). These variables and respiratory rate, FiO2, ventilator settings (ie, ventilatory mode, tidal volume, minute ventilation and positive end expiratory pressure) and cardiac output were recorded in the intensive care unit flowsheet every 30 min. Cardiac output was measured with the use of a pulmonary artery catheter. Serum sodium and osmolarity before and after treatment were also measured. As the duration of ICP reduction usually is maintained for 60–120 min after bolus administration of both mannitol23 and HTS,7 24 ICP and other physiological variables were averaged at 30, 60 and 120 min after each bolus administration. Treatment baseline was defined as the average of all values obtained during the 120 min before a bolus was administered. At each time point, means for each variable were measured for each bolus of mannitol and HTS.

Statistical analysis

As patients received a variable number of boluses of mannitol and HTS, one grand average of treatment was calculated for each individual patient. For each variable, differences between mannitol and HTS treatments at all time points were then analysed with ANOVA for repeated measures. The JMP starter software (SAS Institute Inc, Cary, North Carolina, USA) was used for data analysis. For all analyses a p value <0.05 was considered to be statistically significant.


Patient characteristics

Twelve consecutive patients with severe TBI (nine men and three women, mean age 36 (16) years, median admission GCS score 3 (range 3–8)) were studied. Baseline clinical and demographic characteristics are shown in table 1. One-third of patients died.

Table 1 Patient clinical characteristics

Effect of osmotherapy on PbtO2

A total of 42 episodes of intracranial hypertension treated with mannitol (n = 28 boluses) or HTS (n = 14 boluses) were analysed. The median number of boluses analysed per patient was 2 (range 1–4) for mannitol and 1 (1–2) for HTS. The median time interval between mannitol and HTS administration was 8.6 (interquartile range 4.3–16.7) h. Effect of HTS and mannitol on brain oxygen is shown in fig 1. After HTS, PbtO2 increased from a mean pretreatment value of 28.3 (13.8) mm Hg to 34.9 (18.2) mm Hg at 30 min, 37.0 (17.6) mm Hg at 60 min and 41.4 (17.7) mm Hg at 120 min (all p<0.01). In contrast, mannitol was associated with a non-significant decrease in PbtO2 from the pretreatment value (30.4 (11.4) mm Hg) at 30 min (28.7 (13.5) mm Hg), 60 min (28.4 (10.6) mm Hg) and 120 min (27.5 (9.9) mm Hg). Compared with mannitol, HTS treatment was associated with higher levels of PbtO2 at all times analysed. In particular, brain tissue oxygen after HTS was significantly greater at 60 and 120 min than after mannitol (see fig 1).

Figure 1

Line graph illustrating mean (SD) brain tissue oxygen pressure (PbtO2) at baseline (time 0) and at 30, 60 and 120 min after hypertonic saline and mannitol bolus administration. *p<0.05, **p<0.01 for comparisons between the two treatments.

Effect of osmotherapy on other cerebral and systemic variables

Mean pretreatment values for ICP, MAP, CPP, CVP and cardiac output were similar for mannitol and HTS treatments. Baseline and post-osmotherapy PaO2/FiO2 ratio, respiratory rate and ventilator settings (including FiO2) were similar in the two treatment groups. Arterial blood gas analysis was not performed during osmotherapy: however, SaO2, FiO2 and ventilator settings were stable during all interventions.

Mannitol and HTS were both associated with a significant ICP reduction. However, at 60 and 120 min, HTS treatment was associated with lower ICP and higher CPP than mannitol (table 2). The decrease in ICP and PbtO2 increase did not demonstrate a significant correlation. In addition, HTS bolus administration was associated with an increase in cardiac output that was more significant than mannitol at all time points analysed. MAP and CVP did not differ significantly between treatment groups at each time point.

Table 2 Physiological variables before and after a mannitol or hypertonic saline bolus for elevated intracranial pressure

Baseline serum sodium and osmolarity were similar before mannitol or HTS administration. At the end of the study period, HTS treatment was associated with higher serum sodium (141 (6) before vs 149 (6) mmol/l after; p<0.01). Osmolarity did not differ significantly between the two treatments. After HTS treatment, hypernatraemia (serum sodium >155 mmol/l) was observed in three patients. No other complications (eg, pulmonary oedema, renal failure) associated with osmotherapy were observed.


We analysed 12 consecutive patients who received osmotherapy with both mannitol (28 treatments) and HTS (14 treatments) for refractory intracranial hypertension after severe TBI. We observed that: (1) HTS was associated with a significant improvement in PbtO2; (2) HTS treatment was also associated with both an effective reduction of ICP and a significant improvement in cardiac output; and (3) compared with mannitol, HTS had a more favourable effect on PbtO2, ICP, CPP and cardiac output.

Hypertonic saline and PbtO2

The positive impact of HTS on PbtO2 was observed within 30 min and was statistically significant at 60 and 120 min. Hypertonic saline solutions may improve PbtO2 by increasing CBF, through multiple complementary mechanisms involving the cerebral macro- and/or microcirculation. Firstly, CBF may increase as ICP decreases, thereby increasing CPP. Consistent with this, the increase in PbtO2 was significant at 60 and 120 min when the beneficial effect of HTS on ICP and CPP levels were most notable. In addition, as the PbtO2 probe is in white matter that appears “normal” on CT it is possible that autoregulation is altered. However, we did not measure CBF or autoregulation and this will need future study. Another possibility is that HTS improved PbtO2 through a relative augmentation of cardiac output which, in turn, may have caused an increase in oxygen delivery and CBF. The effect of HTS on cardiac output has previously been documented in non-neurological25 and neurological critically ill patients.26 For reasons that are poorly understood, intracranial hypertension is associated with reduced systolic function.27 28 It is therefore plausible after severe TBI that HTS augments cardiac output not only by directly increasing cardiac preload but also indirectly by lowering ICP. This hypothesis warrants further study. A final possibility is that HTS improves PbtO2 by improving flow in the cerebral microcirculation, by reducing serum viscosity, by improving endothelial function12 29 and/or through anti-inflammatory and antiapoptotic properties,30 all of which improve CBF and brain oxygen delivery.11 20 21 31 Our findings are consistent with others in patients with acute brain injury22 and suggest that hypertonic saline solutions may have beneficial effects on brain oxygenation in patients with severe TBI and intracranial hypertension.

Hypertonic saline and ICP

Although the potential beneficial effects of HTS on brain pathophysiology were first observed in 1919, mannitol is frequently administered as firstline osmotherapy for intracranial hypertension. However, in recent years, there has been resurgent interest in the use of HTS. Recent small clinical series suggest that HTS is an effective agent to treat cerebral oedema and elevated ICP.5 6 Our results are consistent with these findings.

Comparison between HTS and mannitol

Few clinical studies have compared HTS with mannitol and there is currently insufficient data to support the use of one over the other. In experimental intracerebral haemorrhage, Qureshi et al found that none of these treatments had a significant influence on CBF or cerebral metabolism.23 Recent clinical studies in TBI patients showed that mannitol and HTS had similar effects on elevated ICP and brain oxygen, at least when given in parallel and in equimolar doses.4 In contrast with these clinical observations, other studies found HTS to be superior to mannitol in reducing ICP in patients with severe TBI.5 6 32 33 Furthermore, in animal models of intracranial hypertension, HTS also appears to provide better neuroprotection than mannitol.30 3436 Therefore, whether HTS may be superior to mannitol for control of intracranial hypertension is still a controversial issue. As we did not compare mannitol and HTS in a parallel or randomised fashion, and the treatments were not administered in equimolar doses, our study cannot provide a definitive answer to this question. Rather, our results suggest that HTS may have a more favourable effect on PbtO2 and cerebral and systemic haemodynamics than mannitol when administered as a second tier therapy for elevated ICP refractory to mannitol in patients with severe TBI. Our data are also consistent with previous observations in patients with severe TBI that suggest that mannitol has a limited effect on PbtO212 19 but this warrants further clinical investigation.

Study limitations

Our study has several potential limitations. First, HTS was administered after mannitol in all patients. Therefore, our data only suggest that HTS improves PbtO2 and systemic haemodynamics when administered as a secondline osmotherapy in patients with severe TBI and recurrent episodes of intracranial hypertension. In this context (ie, after mannitol), HTS had more favourable effects on brain oxygen and ICP than the previous mannitol treatment. Second, the observed physiological changes after HTS administration may represent the cumulative effect of mannitol and HTS rather than HTS alone. However, pretreatment serum osmolarity and serum sodium were similar in the mannitol and HTS groups, and the time interval between each treatment was relatively long (median 8.6 h), suggesting that a cumulative effect is less likely. Third, mannitol and hypertonic saline were not given at equimolar doses, and we cannot exclude the fact that the higher osmolarity of HTS may, at least in part, explain some of the observed benefits on brain oxygen. Additional studies are needed to analyse whether mannitol and HTS may have comparable beneficial effects when administered in equimolar doses.4 Fourth, data were obtained from only 12 patients and therefore the results should be considered preliminary. However, a total of 42 episodes of elevated ICP were analysed and the effects of 28 boluses of mannitol and 14 boluses of HTS were studied. Fifth, the retrospective nature of the analysis may have introduced bias. However, patients were treated in a standardised fashion and the data were collected prospectively. Sixth, patients were examined at different time points after the initial injury, and variations in CBF and PbtO2 may therefore have occurred over time. Each patient however served as his/her own internal control, and this may have partially reduced the potential influence of time. Seventh, given that MAP was comparable between mannitol and HTS therapy, a possible additional explanation for the observed HTS associated increase in cardiac output may be a decrease in peripheral resistance or, alternatively, a difference in vasopressor dose between mannitol and HTS. These data were not available in our dataset, and we are unable to more precisely address these issues. However, we wish to point out that vasopressors were only used at low dose to maintain adequate CPP and none of the patients included in the study had low cardiac output (cardiac output ranged from ∼6 to 8 l/min) or circulatory shock requiring vasopressors. Finally, we examined the effect of only a single dose of 7.5% HTS and so could not observe any dose dependent effect to confirm our findings.

Despite these study limitations, our findings suggest that hypertonic saline solutions may significantly improve brain oxygen and systemic haemodynamics in patients with intracranial hypertension after severe TBI. Randomised clinical studies are needed to confirm our findings and to examine whether hypertonic saline may be superior to mannitol for the treatment of elevated ICP when both treatments are given in equimolar doses and weight adapted.


Our data suggest that osmotherapy with 7.5% hypertonic saline, when given as a second tier therapy for elevated ICP, is associated with a significant improvement in brain oxygen, CPP and cardiac output in patients with severe TBI and intracranial hypertension refractory to previous mannitol administration. In this context, hypertonic saline also appears to have a more beneficial effect on cerebral and systemic haemodynamics than mannitol.


The authors thank Professor François Feihl for the careful review of statistical analysis.



  • This work was performed at the Neurointensive Care Unit, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, USA.

  • Funding: Supported by Research Grants from the SICPA Foundation, Switzerland (to MO and EC), the Swiss National Science Foundation, Grant PBLAB-119620 (EC), the Integra Foundation (PDL) and the Mary Elisabeth Groff Surgical and Medical Research Trust (PDL).

  • Competing interests: None.

  • Ethics approval: The study was approved by the Institutional Review Board, Hospital of the University of Pennsylvania, Philadelphia, USA.