To date, a relatively small amount of homogenous data has been published regarding the use of TTM with an automated feedback-controlled device for managing temperature in severe TBI. This underlines the importance of consensus discussion in identifying areas of uncertainty where evidence is lacking, and to stimulate harmonisation of processes of care across different settings. To guide discussions, clinical terms were defined with the values as shown in table 2.
Table 2: Definitions applied during the Delphi process.
Clinical term
|
Definition
|
Mild hypothermia
|
Core temperature 34.0°C–36.0°C
|
Therapeutic hypothermia
|
Core temperature <36.0°C
|
Controlled normothermia
|
Core temperature 36.0°C–37.5°C
|
Fever
|
Core temperature >37.5°C
|
Pathophysiology
(i) Temperature measurement and control is an essential aspect of high-quality care in patients with severe TBI
(ii) In patients with impending cerebral herniation, temperature control is essential
As an introduction to the discussions, the group debated the recommendation for temperature measurement and control following severe TBI and, after extensive discussion, concluded that core temperature measurement and control is essential for the provision of high-quality care, especially in patients perceived to be at high risk of secondary brain injury. Noting the phrasing of ‘temperature control’ in the recent guidelines for temperature control following cardiac arrest, 18 the group agreed that as an entry point into high-quality care following TBI, the notion of temperature measurement and control is key, opening the door to the full practice of targeted temperature management. This nuanced phrasing was intended to set the scene for the group’s work, with the specifics of the TTM process such as temperature ranges and duration of control being addressed throughout the remainder of the discussions.
Highlighting the wealth of physiological data available on the management of temperature in stroke and cardiac arrest, the group noted that the guidelines for temperature management in TBI are less specific. Fundamentally, the group agreed that high-quality TBI care does include monitoring temperature and implementing some form of temperature control, recognising its potential role in optimising outcome. The group highlighted the importance of treatment titration based on an individualised risk-benefit assessment and stratification. In particular, it was noted that in patients with exhausted intracranial compensatory reserve and at risk of cerebral herniation or ischaemia – there exists an extreme susceptibility to secondary brain injury precipitated by suboptimal temperature control.
Cerebral herniation is a life-threatening event that requires early diagnosis and prompt management in order to prevent irreversible pathological cascades that can lead to death. 19 Increases in brain temperature have been linked to a linear rise in ICP, with the relationships between temperature, ICP and cerebral perfusion pressure (CPP) becoming more apparent with rapid temperature changes. The impact of temperature on ICP supports the recommendation from the group that temperature control is an essential aspect of care in patients at risk of herniation. 20 The group agreed that while control of ICP and prevention of herniation were important reasons for TTM in TBI, benefits of TTM in the acute phase of TBI also extended to patients without intracranial hypertension.
During the discussions the group highlighted that different pathologies often dictate different patient management. For example, patients in whom fluctuations in ICP are well-tolerated (e.g., patients with high intracranial compliance) will be managed differently to patients with obliterated basal cisterns, obliterated cortical sulci, and midline shift (e.g., intracranial mass effect). In patients with exhausted intracranial volume-buffering reserve, strict control of physiological parameters such as CO2 and temperature, is strongly recommended.
Monitoring
(i) Continuous temperature monitoring is preferable over intermittent temperature measurements in patients with severe TBI
(ii) Monitoring core temperature (e.g., bladder, oesophageal, brain) is strongly recommended over measuring or monitoring superficial temperature (e.g., skin, tympanic) in severe TBI
(iii) When brain temperature monitoring is in place, it is advisable to assess an additional source of core temperature monitoring (i.e. oesophageal, bladder)
The group widely agreed, in line with supporting literature, that continuous temperature monitoring is preferable over intermittent temperature measurements with severe TBI. Intermittent monitoring and recording of temperature can result in large fluctuations in temperature being missed, as highlighted by supporting literature investigating the use of TTM following cardiac arrest, TBI and stroke. 17,21,22
Discussions amongst the group drew attention to the fact that inaccurately measured temperatures can negatively impact patient care and outcome. Several temperature monitoring sites are available for TTM, and the group widely agreed that core temperature measurements, i.e., bladder and oesophageal sites, are strongly preferred over superficial measurements such as those taken at skin and tympanic sites. Following acknowledgement of their limitations, 23 bladder and oesophageal were singled out as favoured core temperature measurements. The group acknowledged the widespread use of oesophageal probes due to their relative ease of insertion and the challenges of finding MRI compatible bladder probes. Confirmation of preference between the two was acknowledged as being beyond the scope of the group due to these nuances. Rectal temperature monitoring was widely regarded as impractical for reasons such as the lag time and a high rate of dislocation. 16,23 Peripheral sites were unanimously deemed to be insufficiently accurate to guide temperature treatment. 16
Some panel members argued that monitoring target organ (i.e. brain) temperature could add a layer of clinical safety, improve pathophysiological understanding and allow selective and individualised titration of treatment (i.e. selective brain cooling). It was, however, agreed by the group that more research is needed into optimum methods for measuring brain temperature and its interpretation from both a clinical and resource-availability perspective. In particular, it was highlighted that temperature thresholds for harm are less well defined for brain temperature than core temperature. When brain temperature monitoring is available and in place, the group advised that core temperature should also be assessed with bladder or oesophageal probes since this is part of routine practice and has been studied to a greater extent than brain temperature. The group noted the importance of having a dual source of temperature monitoring when using automated TTM devices to reduce the risk of probe malfunction and subsequent over or undercooling. 24
After TBI, brain temperature has often been shown to be higher than systemic temperature and can vary independently, with literature noting a difference of as much as 2°C depending on the individual characteristics of brain pathology and/or probe location, making a consistent and accurate link between the two challenging and possibly inaccurate. 25,26 The group highlighted that targeting brain temperature may allow precise titration of treatment dose, including titration of selective brain cooling with brain temperature management technologies, theoretically reducing side effects associated with systemic hypothermia, whilst delivering neuroprotection and brain temperature management. However, it was concluded that further research is needed in this regard and that not enough evidence exists to support practical recommendations.
ICP management
(i) Temperature control is a key component of ICP management in severe TBI
(ii) Controlled normothermia (i.e., target core temperature 36.0–37.5°C) should be included as an addition to the Tier 1 and Tier 2 treatments defined within the SIBICC 2019 guidelines
(iii) Therapeutic hypothermia (i.e., target core temperature ≤36.0°C) should be considered in cases where tier 1 and 2 treatments (as per SIBICC guidance) have failed to control ICP
(iv) If hypothermia is considered to control ICP, target temperature should be managed as close to normothermia as possible
ICP monitoring remains a critical component in the management of severe TBI. 27,28 The group unanimously agreed that temperature control is a key aspect of managing ICP, highlighting that an increase in temperature can lead to an increase in cerebral metabolism and augmented cerebral blood flow, and a simultaneous increase in cerebral blood volume. In cases of exhausted compensatory mechanisms, these factors can precipitate intracranial hypertension, 20 which in turn can have a deleterious effect on overall outcome.
Because there is often no single pathophysiological pathway of ICP elevation, its management is complex. The most recent versions of the Brain Trauma Foundation TBI guidelines do not contain treatment protocols, in part due to a lack of solid evidence around the relative efficacy of available interventions. 27 To address this, the Seattle International Severe Traumatic Brain Injury Consensus Conference (SIBICC) developed a consensus-based practical algorithm for tiered management of severe TBI guided by ICP measurements. 28
One of the most impactful outcomes from this consensus meeting was the acknowledgement of the essential role of temperature control for ICP management in severe TBI, and the recommendation that controlled normothermia (i.e., target core temperature 36.0°C–37.5°C) should be considered in addition to Tier 1 and Tier 2 treatments. The group was keen to harmonise this output with SIBICC by suggesting a more aggressive and specific management with the addition of controlled normothermia in Tiers 1 and 2, adding a layer of clinical safety beyond merely the avoidance of fever over 38.0°C in Tier 0, as shown in Figure 2. In cases when hypothermia is considered (i.e., SIBICC Tier 3), the group recommended that target temperature be managed as close to normothermia as possible, based on an individualised risk-benefit assessment. 29
No consensus was reached on whether hypothermia was a viable temporising strategy in patients with impending cerebral herniation, in patients awaiting hematoma evacuation or decompression, or before consideration of barbiturate coma. Whilst the group acknowledged that therapeutic hypothermia can be effective in reducing ICP, there was no consensus on whether this could be induced rapidly enough in these circumstances, and it was felt that insufficient evidence was available to provide pragmatic recommendations on its indication in these extreme clinical circumstances.
Whilst the majority of experts indicated 35.0°C as the lowest target temperature to be considered in these circumstances, no consensus was reached. The discussion highlighted that insufficient evidence exists to support practical recommendations and highlighted the importance of an individualised risk-benefit assessment. It was also noted that centres might have a varying degree of familiarity with different therapeutic options, including ease of access to neurosurgical options (i.e. ventricular drainage, decompression) and this may have an impact on clinician preference for hypothermia as a temporising therapeutic modality.
The group also discussed the indication of barbiturates in the context of ICP control following severe TBI, not reaching consensus on whether therapeutic hypothermia should be attempted before considering barbiturates. The group noted that both barbiturate-induced burst-suppression and therapeutic hypothermia have distinctive side effects and concluded that no recommendations for standard clinical practice could be made beyond what was already stated in SIBICC guidance.
Fever
(i) Neurogenic fever (core temperature >37.5°C) driven by neurological dysregulation in the absence of sepsis or a clinically significant systemic inflammatory process is relatively common in TBI, and it should be promptly detected and treated (i.e., with controlled normothermia targeting 36.0°C to 37.5°C), irrespective of ICP level
(ii) Controlled normothermia should be considered when pyrexia is secondary to sepsis or inflammatory processes, and when the patient is perceived to be at risk of secondary brain injury, especially in the acute phase of TBI
(iii) Uncontrolled fever (neurogenic or secondary to inflammation or infection) can precipitate secondary brain injury in patients with severe TBI
It was widely agreed that neurogenic fever, defined here as core temperature >37.5°C driven by neurological dysregulation in the absence of sepsis or a clinically significant inflammatory process is common in intensive care and it has been found to be associated with an increased risk of complications and unfavourable outcome. 9,14,15 In the setting of neurogenic fever developing in comatose patients with acute traumatic encephalopathies, controlled normothermia targeting 36.0– 37.5°C was recommended in tier 1 and 2 of the ICP management algorithm.
Correctly differentiating central fever against fever of infectious origin is both challenging and clinically important due to the impact of failing to identify a treatable condition, the negative consequences of antibiotic overuse, and the detrimental effect of hyperthermia on brain-injured patients. 17,30,31 However, the group noted that physiological processes such as brain metabolic rate of oxygen, CO2 control, PbtO2 and ICP are directly related to temperature, and that the deleterious effects and likelihood of secondary injury may occur irrespective of whether temperature is raised due to infection or impaired thermoregulation. This therefore highlights the need for acute management of temperature regardless of the source of the pyrexia, although added focus must be placed on the management of nuanced patient characteristics such as those with severe TBI with impending herniation and/or obliterated basal cisterns, as opposed those with low ICP and preserved intracranial compliance.
In line with current research, 9,11,32 it was agreed that the development of fever is common in TBI cases, and that it can precipitate secondary brain injury and adversely affect patient outcome. It is therefore of utmost importance to prevent or promptly treat fever when detected. The group agreed that while some degree of controlled pyrexia may be allowed during the subacute phase of disease, ‘uncontrolled’ fever requires urgent management in the acute phase as long as the patient is still perceived to be at significant risk of secondary brain injury.
(i) Fever control is recommended in patients with severe TBI who have seizures or are perceived to be at high risk of seizures
(ii) In patients with severe TBI who are sedated and ventilated, controlled normothermia, irrespective of ICP, should be initiated reactively when fever is detected
(iii) When neurogenic fever is detected in TBI cases, controlled normothermia should be continued for as long as the brain remains at risk of secondary brain damage
The group strongly recommended that fever control and controlled normothermia are of particular relevance in patients perceived to be at high risk of seizures and, more in general, secondary brain injury. The assessment of whether an individual patient should be considered ‘at risk of seizures’ or ‘at risk of secondary brain injury’ remains the responsibility of the managing physician. The group defined risk factors for seizures as a history of seizures, the presence of temporal contusions or depressed skull fractures. Features associated with a higher ‘risk of secondary brain injury’ included labile ICP, obliterated basal cisterns, midline shift or subfalcine herniation, and other signs of exhausted intracranial volume buffering reserve. While no consensus was reached on a specific temperature range to target during controlled normothermia, the group agreed that the reactive initiation of temperature control was important in sedated and ventilated TBI patients, with agreement on a pragmatic setting of a target core temperature range of 36.0–37.5°C to accommodate expected fluctuations of +/- 0.5°C while avoiding spikes over 38.0°C. 28
TTM induction
(i) It is recommended that the rapid induction of hypothermia in traumatic brain injury cases should be achieved with automated feedback-controlled temperature management devices
In line with previous research, 17 the group widely agreed on the reactive use of an automated feedback-controlled device for the application of optimal TTM. The TTM process can be divided into three phases: induction, maintenance, and rewarming. 9,16 As explained in existing literature, varying availability of devices and financial aspects may dictate choice, and while non-automated methods of temperature control are cheaper and easier to apply, the level of control offered is poor and their use should be limited to the induction phase, as adjuncts to automated devices. 17,33 The application of therapeutic hypothermia requires constant monitoring of core body temperature in order to achieve an accurate target temperature during induction to prevent overcooling, to assess variations during the maintenance phase, and to ensure a steady, controlled rewarming phase. 16
There was no agreed recommendation from the group as to whether ICUs should stock readily available ice-cold NaCl solutions of different concentrations for the management of ICP crises, citing a lack of clear evidence to draw upon. The group did however highlight the fact that the rapid infusion of ice-cold saline is an inexpensive and readily available option for lowering core body temperature, 9 with the rapidity of response to ice-cold infusions being regarded as a valuable aspect of TTM induction.
TTM maintenance
(i) An automated feedback-controlled TTM device that enables precise temperature control is desirable for the initiation of TTM and maintenance at target temperature in patients with severe TBI
(ii) The maximum temperature variation that a patient should experience during
normothermia is less than or equal to +/- 0.5°C per hour and ≤1°C per 24-hour period
(iii) When hypothermia is indicated, treatment should be continued for as long as the brain is considered to be at risk of secondary brain injury
Automated feedback-controlled devices for TTM are powerful tools, encouraging the delivery of quality care and aiming to improve neurological outcome, 13,17 minimising the chances of temperature variability. Temperature variability is the deviation of patient temperature outside of the goal, typically reported as mean deviation or percent of time outside of target. 9 The group noted that there is a level of pragmatism to be adopted in TTM maintenance, discussing that while more time spent in fever can negatively impact neurological outcome, fluctuations in temperature may also affect outcome, 17 and consensus was reached on the importance of maintaining temperature at as consistent a level as possible with the group settling on a fluctuation range of less than or equal to +/- 0.5°C per hour and ≤1°C per 24-h period. In instances where an automated feedback-controlled device is not available, the group noted the importance of increased staff awareness of patient status to ensure fluctuations outside of this range are appropriately managed. The group highlighted that a dedicated protocol for sedation, analgesia and shivering management might be helpful to ensure consistent application of optimal TTM.
The group agreed that when indicated, hypothermia should be continued for as long as the individual practitioner considers the brain to be at risk of secondary injury. These considerations were supported with a suggestion that it should be maintained for as short a time as possible.
TTM rewarming
(i) Obtaining an interval scan and/or an alternative assessment of intracranial compliance, in addition to the absolute number of ICP, is recommended before rewarming
(ii) Rebound hyperthermia should be prevented whenever possible or promptly treated in cases when the brain is perceived to be at risk of secondary brain injury
In cases in which the patient is being rewarmed from therapeutic hypothermia (core temperature lower than 36.0°C), the group agreed that once ICP has been maintained within controlled limits and de-escalation of treatment intensity is considered, it is sensible to ensure the patient has sufficient intracranial volume buffering reserve through the use of an interval scan and/or an alternative measure of intracranial compliance, before commencing the rewarming process. The group also noted the high prevalence and potential risks associated with rebound hyperthermia when TTM is discontinued following therapeutic hypothermia, highlighting the importance of continued vigilance and careful temperature control in the rewarming phase.
Whilst no consensus was reached on recommended rewarming rates, the group agreed that controlled rewarming with an automated feedback-controlled device may reduce the risk of rapid temperature variations and rebound pyrexia that can precipitate secondary brain injury and compromise care. 16,33 The group highlighted how controlled rewarming may improve the ability of clinicians to more effectively control important inter-dependent clinical variables such as PaCO2, ventilation settings and depth of sedation.
TTM for shivering
(i) It is important to assess, document and manage shivering in severe TBI patients
(ii) Whenever ICP is labile and shivering is detected, neuromuscular blockers should be considered after ensuring appropriate depth of sedation
(iii) In self-ventilating patients in the subacute phase of severe TBI, an individualised risk-benefit assessment should be undertaken regarding the strict indications of controlled normothermia
(iv) Permissive hyperthermia should be considered in cases where risk of secondary brain injury resulting from pyrexia is thought to be low, and when shivering cannot be controlled with first line treatments such as NSAIDs, opiates, magnesium or counter warming
In line with current literature, it was widely agreed that shivering should be managed in patients following severe TBI. Shivering can reduce brain tissue oxygenation leading to cerebral metabolic stress, which may therefore negate the neuroprotective benefits of TTM. 9,34–36
Titration of sedation and the use of neuromuscular blocking agents provides intensivists with readily available and effective options for shivering control in critically ill patients. 37 To ensure appropriate and effective use however, treating staff must be aware of the nuances of selecting the correct agent, monitoring the depth of neuromuscular blockade, and ensuring adequate skeletal muscle recovery once therapy with neuromuscular blockers has ceased. In cases of shivering when ICP is labile, the group agreed in line with current literature that ensuring depth of sedation before administering neuromuscular blockers is of utmost importance. 37, 38 When using pharmacologic agents for shivering management, treating staff must consider potential pharmacokinetic and pharmacodynamic variation and monitor for efficacy (i.e. shivering control) and safety (i.e. adverse events and drug-drug interactions). 9
The group agreed that in patients who are perceived to be at relatively lower risk of secondary brain injury (i.e. self-ventilating patients in the sub-acute phase of severe TBI), permissive hyperthermia may be considered over TTM, especially if the latter therapeutic option would require sedation or other invasive interventions. The group agreed that an individualised risk-benefit assessment should ultimately be undertaken before commencing controlled normothermia in such patients.
Auditing
(i) ‘Time within target range’, ‘burden of fever’ and similar metrics can be considered as indicators of quality of temperature management
‘Time within target range’ and ‘burden of fever’ were considered by the group to be appropriate metrics of quality temperature management. It was widely acknowledged that these metrics should be weighed by patient length of stay and/or duration of monitoring for appropriate statistical interpretation. The group was also careful to note that the administrative burden on physicians is already high and acknowledged the fact that some centres may not have access to electronic patient data management systems, so it was agreed that it was unrealistic for this group to issue prescriptive recommendations on auditing practices. In light of the high heterogeneity across centres, 9 here the group were keen to clarify that wherever possible, documenting metrics such as ‘time within target range’ and ‘burden of fever’ may improve their ability to deliver data-driven service improvement and temperature control.
Summary
This consensus review was undertaken to evaluate current evidence on the application of TTM in the management of severe TBI in a critical care setting, and to develop a set of practical recommendations to address identified gaps in current published evidence.
As highlighted by the SIBICC 2020 group, the gap between published evidence and management protocols is bridged by expert opinion. 39 The optimal method for the provision of high-quality TTM remains unknown, and barriers to its consistent implementation include the lack of evidence-based treatment protocols, knowledge deficiencies, limited access to equipment, lack of financial resources and staff workload. This document aims to address key practice gaps and optimise patient care through multimodal assessment following TBI.
Strengths and limitations
The Delphi process has a number of strengths. Participants are able to reconsider their views in light of the evolving discussions, allowing for an element of reflection that isn’t regularly seen in other studies involving a single time point such as interviews or focus groups. 40 The element of anonymity offered to the panellists in the survey rounds avoids group conformity and promotes honesty, and the controlled and iterative discussions offer a flexible approach to gathering expert viewpoints on the set research questions. The Delphi method is an iterative process allowing the anonymous inclusion of a number of individuals across diverse locations and areas of expertise and avoiding dominance by any one individual. It uses a systematic progression of repeated rounds of voting and is an effective process for determining expert group consensus where there is little or no definitive evidence and where opinion is important. 41,42 The modified Delphi approach used here combined the early flow of structured information and submission of anonymous responses with the (hybrid) face-to-face discussion and further voting to gain consensus (or establish lack thereof) and expert insight into usual practice regarding non-pharmacological TTM with an automated feedback-controlled device. As cited in existing literature however, 13,17 the Delphi process has limitations. The process is vulnerable to drop-outs and technical issues, with the online voting process during our meeting seeing some participants unable to cast their votes on a number of questions, leading to the need for a final anonymous survey round. The group opinions during the meeting may have been impacted by social bias, and the voices across the in-person and online participants may not have been equally heard, highlighting a potential need to ensure consistency in attendance in the same format in future panel meetings.
This report has been developed by an expert panel comprised of specialists in neuro-critical care experienced in the management of severe TBI, therefore the recommendations focus on patients managed in a critical care environment. An individualised risk-benefit assessment should be undertaken for each domain to accommodate the high levels of heterogeneity seen across TBI patients, local practice settings, staff training and equipment availability. 9