A Critical Point of View in the Management of Intracranial Hypertension: Are All Therapeutic Tools Evidence Based?

31 A Critical Point of View in the Management of Intracranial Hypertension: Are All Therapeutic Tools Evidence Based?

Thomas Lescot ¹, Lamine Abdennour ¹, Louis Puybasset ¹

¹ Neurosurgical Unit, Department of Anesthesiology and Critical Care Pitié-Salpêtrière Hospital, Assistance Publique, Hôpitaux de Paris, Université Pierre et Marie Curie, Paris, France

31.1 Introduction

Most TBI patients with cerebral lesions on cerebral computed tomography (CT), such as hematomas, swelling, contusions or herniation, will develop ICH in the days ensuing injury. A recent observational study showed that the mean intracranial pressure (ICP) will peak within the first 3 days after injury in half of such patients and after 5 days in 25% [1]. An uncontrolled rise in ICP is probably the most common cause of death in TBI patients. For these reasons, ICP monitoring is strongly recommended in TBI associated with abnormal cerebral CT findings. Intracranial pressure monitoring allows to continuously monitor ICP, manage cerebral perfusion pressure (CPP), and withdraw cerebrospinal fluid if an external ventricular drainage is inserted. Elevated ICH is strongly associated with poor outcome in patients with severe TBI. Indeed, elevated ICP may lead to decreased CPP, cerebral ischemia and/or herniation. The ICP threshold above which treatment should be initiated is still debated. A large, prospectively collected database study published in 1991 by Marmarou et al. [2] reported a strong correlation between the number of hours with an ICP >20 mmHg and outcome. Similar results have been recently reported by Balesteri et al. in a retrospective analysis involving 429 TBI patients [3].

Does treatment to decrease ICP improve outcome in TBI? Few data exist regarding this question, and no controlled randomised trial is available, but it seems that patients in whom ICP could be controlled have a much better outcome than those with in whom ICP remained uncontrolled [4,5]. More recently, two retrospective studies showed similar results, suggesting an improved outcome with aggressive ICP therapy [6,7].

In the treatment of ICH, one of the key issues to consider is that TBI is more likely a syndrome than a disease. After trauma, the traumatised brain is characterised by a marked pathophysiological heterogeneity: ischemic areas (cytotoxic edema) coexists with areas with blood-brain barrier disruptions (vasogenic edema), contusions, and normal brain parenchyma. The proportion of each of these areas likely depends on the etiology of TBI. Intracranial pressure-lowering therapies may have some systemic or cerebral differential effect according to the TBI presentation. For example, one can consider that in blood-brain barrier disruption, treatment such as hypertension-induced high CPP and osmotherapy could worsen contusional areas of the traumatised brain.

31.2 Sedation as a Therapy

Sedatives and anesthetics are commonly given to prevent agitation, minimize noxious stimuli, and adapt TBI patients to mechanical ventilation. Benzodiazepines, especially midazolam, are administered in combination with morphine or derivatives such as sufentanil. While they are particularly suitable in the neurological intensive care unit, they are unable to depress brain electrical activity even at high doses, present a long duration of action and a plateau effect.

Propofol administration has been proposed as an alternative in TBI patients. Propofol decreases ICP by reducing brain metabolism [8], which explains its potential neuroprotective effect. Only one double-blind controlled randomised trial was conducted to compare propofol infusion with another sedative regimen in TBI patients. Kelly et al. [9] evaluated the clinical safety and multiple clinical and radiological endpoints in 23 patients in the propofol-morphine group and 19 patients in the morphine sulphate group. The authors found a lower incidence of ICH at day 3 in the propofol group, without any difference in adverse effects between groups. The combined use of midazolam and propofol allows the control of ICH and the possibility to obtain burst-suppression at high doses. Such a combined strategy reduces the use of barbiturates. However, the potential risk with using propofol is the feared propofol infusion syndrome. This syndrome, characterized by multiorgan failure, has a high incidence in sepsis or septic shock, which are therefore contraindications to propofol administration. It is mandatory to stop propofol in case of metabolic acidosis (with or without lactates), hyperkalemia, renal insufficiency, rhabdomyolysis or triglyceride level >5 mmol/l [10]. A daily dosage of blood triglycerides is a common practice in our unit. Hypertriglyceridemia is a warning symptom in this context.

The occurrence of refractory brain hypertension during the first days following the onset of head trauma often requires therapeutic escalation with such therapeutics as neuromuscular blockers, hypothermia and thiopental, all of which have a high potential to deteriorate lung status further through mechanical effects or direct immunosuppressive effects. Barbiturate therapy reduces cerebral blood flow (CBF) and cerebral metabolic rate of oxygen (CMRO₂) and it increases cerebral vascular resistance in patients with severe head injury [11]. Barbiturates, particularly pentobarbital and thiopental sodium, have been used effectively in brain-injured patients with increased ICP refractory to sodium management, cerebrospinal fluid drainage, and moderate hyperventilation. The first published series of head-injured patients receiving this therapy in 1979 [12] found that high-dose barbiturate administration reduced elevated ICP in most patients refractory to a rigorous regimen of ICP management. Several studies in the 1970s suggested that barbiturate therapy had a major impact on the prognosis of patients with intractable ICH. Marshall et al. [12], for example, found that 19 of 25 (76%) patients with severe, intractable ICH responded to treatment with high-dose pentobarbitone therapy and that 10 of the 19 responders made a good functional recovery. The authors concluded that barbiturate therapy significantly improved clinical outcome on the assumption that their patients would have otherwise died.

Other studies showed that high-dose barbiturate, when added to a routine of conventional management, could be useful for the reversal of ICP. Barbiturates seem to provide the brain with some protection against anoxia, ischemia, and cerebral edema by decreasing the brain’s oxygen consumption, reducing CBF, and stabilizing the membranes. However, barbiturates have a number of serious adverse effects. Cordato et al. [13] reported that prolonged barbiturate infusion induces respiratory complications in 76%, lung infection in 55%, arterial hypotension in 58%, hypokalemia in 82%, hepatic dysfunction in 87%, and renal dysfunction in 47% of the treated patients. Continuous barbiturate infusion is also known to produce immunosuppression by inhibiting lymphocyte function [14], affecting neutrophil function and depressing humoral immune response through a decrease in immunoglobulin production [15]. The use of barbiturates is, by itself, a statistical predictor of an increased risk of pneumonia [11].

31.3 Hyperventilation

The simplest way to decrease ICP is to reduce arterial partial pressure of carbon dioxide (PaCO₂). Reduction in ICP is secondary to a decrease in cerebral blood volume. The impact of a sudden change in blood volume on ICP depends on brain compliance and the position of the patient on the Langfitt curve. Accordingly, a given change in PaCO₂ and cerebral blood volume will markedly decrease ICP in a non-compliant brain and induce nearly no change in pressure in a compliant brain. Such a decrease in cerebral blood volume also occurs when the mean arterial pressure is increased. However, both therapies have opposite effects on CBF.

Cerebral arteries respond to highly localized perivascular alterations of PaCO₂ and pH. During chronic hypocapnia, the CBF remains constant, suggesting that carbon dioxide itself does not control cerebral vascular cerebral tone. It is likely that local pH rather than local carbon dioxide is the mediator of tone regulation. The mediator cascade that links extracellular pH to cerebral vascular tone is complex and interrelated, the final mediator being intracellular calcium. Most studies report a change in global CBF of 1 to 2 ml/100 g^-1 min^-1 for each mmHg change in PaCO_2. Acutely reducing PaCO₂ to 25-30 mmHg decreases the global CBF by 40 to 50%.

Payen et al. [16] studied by contrast-enhanced magnetic resonance imaging the relationship between the change in ventilation and CBF in rats. The authors demonstrated that hypercapnia results in a significant increase in CBF in the brain-studied regions, although there was no direct correlation between the change in cerebral blood volume and flow. Carmona Suazo et al. [17] investigated the effect of hyperventilation on cerebral oxygenation by means of continuous monitoring of brain tissue oxygen pressure (PbrO₂). Acute hypocapnia significantly decreased PbrO₂, indicating the risk of secondary ischemic damage during hyperventilation in severely head-injured patients. These conclusions agreed with those of Imberti et al. [18] who observed similar results, clearly suggesting that hyperventilation can compromise cerebral oxygenation. The most recent guidelines of the Joint Committee on Trauma and Critical Care of the American Association of Neurologic Surgeons indicate that aggressive or prophylactic hyperventilation must be avoided in patients with severe head trauma [19].

31.4 Drainage of Cerebrospinal Fluid

Drainage of cerebrospinal fluid (CSF) is frequently performed in TBI patients. The external ventricular drain (EVD) connected to an external strain gauge allows to continuously measure ICP and to calculate CPP (when the drainage line is externally clamped), or to withdraw CSF (when the drainage line is open). In low cerebral compliance due to cerebral edema, the evacuation of a few ml of CSF may be sufficient to decrease ICP dramatically. Kerr et al. showed that a 3-ml withdrawal of CSF resulted, on average, in a 10% decrease in ICP and a 2% increase in CPP, which were sustained for 10 minutes [20]. This therapy is simple, cost effective and overrides the often serious systemic complications related to drug or physical therapies, especially those induced by barbiturates and hypothermia. Drain placement might be technically difficult, and can be complicated by cerebral contusion or ventriculitis [21]. Two studies have shown that the difference in pressure gradient induced by intracranial CSF drainage facilitates the transfer of cerebral edema to the ventricles, thus improving its clearance [22]. Continuous drainage of CSF against a zero pressure via an EVD, combined with continuous ICP monitoring with an intra-parenchymal catheter, is a common practice in our unit. This technique likely increases the volume of drained CSF as compared to discontinuous drainage, although there are no hard data to support this clinical observation.

31.5 Osmotherapy

In TBI, increased ICP is most often due to cerebral edema. Hyperosmolar agents are widely used to control edema formation after TBI. There is abundant literature supporting the use of mannitol to decrease ICP, increase CPP, and enhance CBF (23) without affecting cerebral oxygenation [24]. Hypertonic saline used at various concentrations (from 3% to 23.4 %) has been consistently shown to decrease ICP and cerebral water content in TBI patients [25]. Hyperosmolar agents were shown to decrease water content in non-traumatised brain tissue by osmotic mobilisation of water across an intact blood-brain barrier (BBB). However, there is the possibility of an opposite effect of the osmotic agent in the areas of a disrupted BBB. This was observed by Saltarini et al. [26] using magnetic resonance spectroscopy to assess cerebral water content in a patient with refractory ICH. One recent CT-based analysis showed an increase in contusion volume after hypertonic saline infusion, suggesting a leak of electrolyte and plasma from the extracellular compartment into the contused area through a disrupted BBB [27]. These results argue for the differential effects of hypertonic saline according to the state of the BBB in various brain areas. As a result, it could be proposed to reserve the use of osmotic agents for patients presenting a small volume of contusion.

31.6 Cerebral Perfusion Pressure

Management of cerebral perfusion pressure (CPP) in TBI is still debated. In 1995, Rosner et al. developed a management paradigm that focused on maintaining CPP >70 mmHg. This strategy, based on BBB integrity and preserved cerebral autoregulation, seemed to improve outcome in an uncontrolled series of TBI patients [28]. An alternate approach has been described and popularized by the Lund group [29]. The Lund theory hypothesises that the BBB is mainly disrupted after head trauma and that cerebral autoregulation is lost. Increasing CPP or using osmotic agents is therefore regarded as a potentially dangerous therapy which can increase edema formation. The Lund protocol therapy has two main goals: 1) to reduce or prevent an increase in ICP (ICP-targeted goal); and 2) to improve perfusion and oxygenation around contusions (perfusion-targeted goal). There is current evidence to support the concept that hypotension is deleterious for the traumatised brain. In contrast, increasing CPP up to 70 mmHg was associated with an increased incidence of pulmonary failure [30].

A recent prospective observational study compared the 6-month outcome of two groups of TBI patients: one group (n=67) treated in Sweden with the aim of keeping ICP <20 mmHg and CPP >60 mmHg, and a second group (n=64) treated in Scotland with a Rosner-derived strategy protocol in which the goal was to reach a CPP of at least 70 mmHg while keeping ICP <25 mmHg [31]. The authors noted that CPP management therapy seemed to be more efficacious in the patients with intact cerebral autoregulation. In contrast, patients with loss of cerebral autoregulation appeared to benefit from Lund CPP management. This discrepancy could be linked to the presence of contusions: preserved autoregulation is more frequently observed in patients with few contusions [32]. Taken together, the results suggest that patients with numerous contusions (BBB mainly disrupted and loss of autoregulation) would benefit from a Lund-based therapy. Conversely, in patients with only a few contused areas, a CPP-targeted therapy associated with osmotherapy would be logical from a pathophysiological point of view.

31.7 Hypothermia

Prophylactic hypothermia is effective in improving outcome in cardiac arrest [33]; however, it remains controversial in other indications such as brain injury. Nevertheless, the results of different studies, and meta-analyses in particular [34-37], did not demonstrate that hypothermia was associated with a reduction of mortality in TBI. However, the systematic use of hypothermia, independently of ICP, may have underestimated its usefulness. Polderman et al. assessed the efficacy of hypothermia therapy (32-34°C) as part of a step-up protocol to detect and treat side effects [35]. They found that artificial cooling could improve survival and neurological outcome, especially in the patient group with low Glasgow Coma Scale (GCS) scores (GCS=5-6). According to these results, it seems logical to reserve hypothermia for patients with increased ICP. Outcome after hypothermia can be optimised when proper indications, techniques for implementation, as well as management procedures and their enforcement are followed. A rigorous and effective application cannot be conceived outside specialized units with experienced teams. After a minimum duration of 24 hours and after controlled ICP, very gradual warming can be started with the possibility to interrupt it when ICP increases [38]. A decrease in brain metabolism associated with anti-inflammatory effects is the main mechanism of action attributed to hypothermia treatment. Kalemia must be strictly controlled during the ascending and descending changes in body temperature. Of note is that the implementation of hypothermia has been recently facilitated with the availability of automatic cooling blankets and specially designed catheters.

31.8 Steroids

Steroids are particularly effective to treat ICH secondary to cerebral tumoral processes and bacterial meningitis. The bulk of available evidence indicates that steroids do not improve outcome or decrease ICP in TBI patients. The randomised controlled CRASH trial [39] involving 10,008 patients showed no effect of corticosteroids on ICP and a higher mortality was observed in the steroid-treated groups. The study demonstrates that systemic corticosteroids are detrimental to TBI patients. However, the inclusion criteria were very broad and the population of TBI patients was extremely heterogeneous.

The use of steroids in TBI patients with severe contusions could be effective in managing ICH owing to their acting on pericontusional edema, which usually markedly increases ICP between 24 and 72 hours after trauma (clinical experience). Their use under these conditions is possible should all other conventional therapies fail to reduce ICP. A treatment duration of 2 to 4 days is often enough in these circumstances. In our unit, methylprednisolone is administered (dose of 120 mg twice a day for 3 days) if ICP increases in a patient with major brain contusions despite maximal medical therapy and temperature control.

31.9 Albumin

There are some experimental findings showing that albumin reduces pericontusional edema. In a model of ischemia-reperfusion in an isolated guinea pig heart model, albumin more effectively prevented fluid extravasation than crystalloid or artificial colloid. This effect may be attributable to an interaction of albumin with the endothelial glycocalyx also present on the endothelial layer of the BBB. In humans, the results of systemic albumin administration are disappointing. A post-hoc analysis of the SAFE study showed deleterious effects of systemic albumin administration [40]. The main flaws of this study were the a posteriori analysis and the slight imbalance between the two groups. Albumin use should be restricted to this sub-population of patients with severe TBI.

31.10 Conclusions

One of the key issues to consider is that TBI is more likely a syndrome than a disease. After trauma, the traumatised brain is characterised by pathophysiological heterogeneity. One and the same therapy for the various different types of brain trauma is not efficient. This could partly explain why systematic use of an increased CPP approach, hypothermia, corticosteroids or albumin failed to demonstrate any positive effect in large multicentre trials. Indeed, the volume of contused brain seems to reflect the state of the BBB and the preservation of cerebral autoregulation, which are the determinants in the choice of therapies. The algorithm in Table 31.1, used in our institution, illustrates therapeutic strategies delineated according to both the side effects of the therapies and the presence or not of a high contusional volume.

Emergency care

1. Mannitol 20% (0.7 to 1.4 g/kg) or hypertonic saline (20%, 40 ml) if pulsatility index >1.4 and/or mydriasis

2. Immediate correction of hemostasis disorders

Only gold members can continue reading. Log In or Register to continue