The incidence and outcomes of hyperacute cardiovascular dysfunction following isolated traumatic brain injury: an observational cohort study

Abstract

Background:

The relationship between early cardiovascular dysfunction (CVD) in isolated traumatic brain injury (iTBI) and outcome has not been fully described. We aimed to (1) determine the prevalence and phenotype of CVD after iTBI in the hyper-acute phase and (2) compare treatment and outcomes in those with CVD vs non-CVD.

Methods:

An observational cohort database study of severe iTBI patients (Head AIS 3+) at a level 1 trauma centre (2008–2019) and physician-led air ambulance service (2019–2020). CV dysfunction was defined as tachycardia or bradycardia, with hypotension. Physiology, laboratory results, 24-hour transfusion, and computer-topography (CT) findings were recorded. Outcomes were 28-day mortality and Glasgow Outcome Score (GOS).

Results:

A total of 168 patients met inclusion criteria, average age 46 years (IQR 30–61), 77% male, median ISS 25 (IQR 17–29) with 51% Head AIS 5. Time from injury to pre-hospital assessment was 31 minutes (IQR 20–42) with 20% demonstrating CVD on initial observations. The CVD group were more shocked (lactate 6.1 (1.7–10.9) vs. 2.4 (1.4–3.3), P < 0.001) and coagulopathic (43% vs. 15%, P = 0.001). There was no difference in Head AIS or CT findings between groups, except frequency of hypoxic ischemic encephalopathy (HIE) (CVD: 21% vs. non-CVD: 1%, P < 0.001). 24-hour transfusion was higher in CVD patients: 3 (0–8) vs. 0 (0–0) units, P< 0.001. Mortality was greater in CVD vs non-CVD iTBI (61% vs. 31%, P = 0.002), but in patients with AIS 5 there was no difference (P = 0.262). One-third of CVD survivors (13/33) were discharged home, and 4/18 patients with recorded GOS had good neurological outcome.

Conclusion:

One in five patients with severe iTBI develop early CVD, associated with increased mortality, coagulopathy, and HIE. However, mortality and neurological outcome is highly variable in those with CVD across the iTBI severity spectrum. Further research is needed to define the pathophysiology and optimal treatment to improve outcomes for this subgroup of iTBI.

Background

Traumatic brain injury (TBI) is recognized as a major global health problem. It is a significant contributor to morbidity and mortality in the young and is the most common cause of death or disability in under 40 years old in the UK.^[1,2] However, its incidence is rising rapidly in those over 50 years old, particularly in high-income countries.^[3] Despite improvements in injury-related mortality over the last 30 years, the same improvements have not been seen in outcomes from TBI.^[4-6] It has been suggested that the critical phase of TBI occurs early after injury in the pre-hospital phase of care and involves changes in cerebral and systemic physiology which impact morbidity and mortality.^[7] It is recognized that the effects of TBI are not isolated to the brain alone; but are multi-system, including the changes to the cardiovascular system that are in part attributed to the neurocardiac axis.

“Shock” following blunt trauma is classically associated with bleeding or loss of sympathetic outflow following spinal cord injury. Until recently, international teaching suggested that shock did not result from TBI alone.^[8] However, cardiovascular dysfunction (CVD) including concurrent tachycardia and hypotension in TBI, has been described in-hospital and is associated with poor outcomes.^[9-11] The mechanisms driving this response are not well understood but are likely to be multifactorial including the role of an apneic phase;^[12] the effects of a systemic catecholamine surge;^[13,14] cardiac myocyte hypercontraction and cell death resulting from local release of noradrenaline (NA);^[15,16] and a systemic inflammatory-mediated response.^[17] It is suggested that early recognition and management of the CV sequalae of TBI may improve clinical outcomes.^[18]

The pre-hospital clinical presentation of shock in undifferentiated patients, provides a diagnostic and therapeutic challenge to clinicians. There is the potential for patients to be incorrectly identified as bleeding, with the subsequent risk of ineffective or inappropriate treatments and unsuitable physiological end points e.g. permissive hypotension. Gavrilovski et al^[19] have previously identified a cohort of patients with TBI that presented with tachycardia or hypotension during the pre-hospital phase, and more recently Partyka et al have described the challenge of distinguishing a patient with hemodynamic compromise associated with TBI from those suffering hemorrhagic shock, in the pre-hospital phase of care^[20]. However, these studies are limited to observational studies with small numbers. There is a lack of granularity to describe the clinical phenomenon of “shocked” isolated TBI patients during the hyper-acute pre-hospital phase, the characteristics of this cohort and the impact of this physiological derangement on treatment and outcome.

The overall objective of this study was to examine the patient characteristics and outcomes of CVD after isolated TBI in the hyper-acute phase following injury from two prospective trauma registries in a mature trauma system. Our aims were to first describe the phenotype of trauma patients who demonstrated CVD with iTBI. Second, compare pre-hospital transfusion and neurosurgical treatment strategies used in patients with CVD and iTBI. Finally, determine the clinical outcomes in iTBI patients with CVD and without CVD.

Methods

This study analyzed harmonized data from two trauma registries:

1. Perpetual observational cohort study at a level 1 trauma centre – Activation of Coagulopathy and Inflammation in Trauma (ACIT-2) study (ISRCTN 12 962 642) using data from 2008 to 2019.

2. Clinical data collected over a one-year period (2019–2020) from a trauma Helicopter Emergency Medicine Service (HEMS).

Study population

Inclusion criteria were as follows: Glasgow Coma Score (GCS) <9 recorded on initial pre-hospital observations, age >16 years and severe isolated TBI according to abbreviated injury score (AIS) (Head AIS >2 and all other body region AIS <3). These inclusion criteria were established from data submitted by the major trauma centre (MTC) to the national Trauma Audit and Research Network (TARN) trauma registry. Patients excluded were those with any other body region AIS >2; those pronounced life extinct on scene; not taken to an MTC; those whose injury and AIS were unknown as conclusive imaging was not conducted in hospital; and any patient whose intracranial injury was not associated with a primary traumatic cause.

The ACIT-2 study includes all patients admitted via a trauma team activation to a level 1 trauma center in London, UK. Exclusion criteria are patients transferred from another hospital, burns >5% total body surface area, more than 120 minutes from time of injury, and those deemed inappropriate for recruitment by an independent clinician. Data is collected from ambulance service pre-hospital patient record forms and in-hospital patient notes. Deidentified data from 1 January 2008 to 1 June 2019 was extracted from the ACIT database.

The HEMS provides pre-hospital trauma care to an urban population of over 9 million and sees approximately 2000 patients per annum. It operates a doctor-paramedic model and during the study period one team was available for dispatch 24 hours a day, by helicopter or rapid response car. During the study period, 1 August 2019–31 July 2020, patients with clinical suspicion of a traumatic head injury and GCS <9 on initial assessment of the HEMS team were identified via a daily review of all cases attended over the previous 24 hours.

Data collection

Data were collected on patient demographics, mechanism of injury and time from injury to hospital. Fall from height was defined as a fall reported as >1 m or >5 steps.^[2] Pre-hospital data was collected from patient records held on the pre-hospital electronic patient database (OnBase, Hyland Software Inc., OH, USA). In-hospital and outcome data were collected from patient hospital records and data submitted to TARN and Intensive Care National Audit and Research Centre by the MTC. Injury characteristics were collected as injury severity score (ISS), AIS, pathology identified on initial computer-topography (CT) scan and physiology on scene and in-hospital (heart rate [HR; beats/min], blood pressure [mmHg], GCS). Pupil equality and reactivity were collected in the HEMS cohort. Each patient was allocated a score for each pathology identified on CT by the reporting radiologist to a maximum of five (extradural hemorrhage, subdural hemorrhage, subarachnoid/intraparenchymal/interventricular hemorrhage, diffuse axonal injury, hypoxic ischemic encephalopathy).

Patients were divided into two groups, those that demonstrated CVD and those that did not (non-CVD) during the hyper-acute pre-hospital phase. CVD was defined as simultaneous HR >100 bpm or <60 bpm, and systolic blood pressure (SBP) <100 mmHg identified on the first pre-hospital observation recorded (ACIT-2), or on screening of all pre-hospital observations prior to pre-hospital emergency anesthesia recorded by the pre-hospital team. In the latter, hypotension was not included if the reading was annotated as “spurious” (or similar) and seemed unlikely given the clinical context; or if there was a non-hypotensive BP reading within two minutes either side of a hypotensive reading. Pre-hospital emergency anesthesia was defined as any drug assisted intubation. Physiological observations after induction were excluded to avoid the well reported and potentially confounding physiological effects of induction drugs on a patient’s hemodynamics.^[21]

Initial blood gas and coagulation screen after arrival in hospital were collected to include pH, lactate, base excess and INR or prothrombin time. Data of interventions collected in the pre-hospital phase were infusion of hypertonic saline (HTS) and transfusion of red blood cells or fresh frozen plasma. HTS is administered in the pre-hospital phase to patients attended by a pre-hospital team meeting the following criteria: clinical suspicion of significant TBI with unilateral or bilateral dilated pupil(s) or progressive bradycardia and hypertension in keeping with a Cushing’s response. In the HEMS cohort, additional data was collected on pre-hospital administration of phenylephrine and pre-hospital emergency anesthesia.

Primary outcome was mortality at 28 days. Secondary outcome data collected was: 24-hour total blood component transfusion, Intensive Care Unit (ICU) length of stay, total length of stay, length of stay of survivors (alive at 28 days), time to death (died within 28 days) and discharge destination (usual residence; nursing home/hospital). Length of stay reported day of admission as day one and included rounded-up bed days. In the HEMS cohort data, Glasgow outcome defined as GOS 4–5 (moderate disability, low disability or good recovery).

Data analysis

Data analysis was conducted using SPSS (IBMM SPSS Statistics, version 27). Categorical data are presented as percentages. Continuous data distribution was analyzed using Shapiro-Wilk tests and presented as median with interquartile range or mean with 95% confidence intervals. Any difference between groups were identified using chi-squared tests for categorical data and Mann–Whitney U test (nonparametric) or Independent samples T-test (parametric) for continuous variables.

Univariant and multivariant logistic regression was conducted on clinical variables predicted to be associated with the independent variable CVD. Only variables conceivably accessible in the pre-hospital phase of care were included (age, sex, pre-hospital GCS, pre-hospital pupil reactivity, and lactate). Significance was defined as P < 0.05.

Ethics

The ACIT-2 study was approved by a research ethics committee (REC reference: 07/Q0603/29). Retrospective analysis of deidentified data did not require further ethics approval. The HEMS cohort met National Institute for Healthcare Research (NIHR, UK) approval for service evaluation and formal ethical approval was therefore not required. This service evaluation was approved by the Research & Clinical Effectiveness Lead at HEMS and registered as a service evaluation at the relevant NHS Hospital Trusts. This study has been reported in line with the STROCCS criteria^[22].

Results

In total, there were 181 patients that met the inclusion criteria, 13 patients did not have complete cardiovascular observations and were excluded leaving 168 patients available for analysis (Fig. 1). Average age was 46 years (IQR 30-61) and 77% were male. Patients were seriously injured with a median ISS 25 (IQR 17-29) and 51% had Head AIS 5. Average time from injury to pre-hospital HEMS team assessment was 31 minutes (IQR 20-42) with 20% of patients showing CV dysfunction (CVD) on initial pre-hospital observations. In the HEMS cohort, 24% of patients demonstrated CVD prior to pre-hospital emergency anesthesia.

Figure 1.

Consort flow diagram.

There was no difference in age, sex, ISS, Head AIS, time from injury to pre-hospital assessment or total pre-hospital time between the CVD and non-CVD groups (Table 1). Distribution of age, mechanism of injury, Head AIS and pathology identified on initial CT in the CVD versus non-CVD groups are demonstrated in Fig. 2A–D. There was no difference in average Head AIS between the groups, but patients with CVD had significantly greater representation at both ends of the AIS spectrum (AIS 3: 33% vs. 16% (P = 0.027) and AIS 6: 9% vs. 0% (P < 0.001) in CVD vs. non-CVD groups respectively) (Fig. 2C). There was no difference in TBI pathology on CT findings between groups, except frequency of hypoxic ischemic encephalopathy (CVD: 21% vs non-CVD: 1%, P < 0.001). On arrival to hospital the CVD group were more shocked with average lactate 6.1 (1.7–10.9) vs. 2.4 (1.4–3.3), P < 0.001) and coagulopathic (INR >1.2: 43% vs. 15%, P = 0.001) than non-CVD patients (Table 1). There was no difference in pre-hospital equality of pupil size between groups, but patients with CVD were more likely to have unreactive pupils on scene (74% vs 26%, P < 0.01).

Table 1.

Patient and injury characteristics

Variable	Non-CVD (n = 135)	CVD (n = 33)
Age	46 (30–60)	40 (30–63)
Sex n (%) male	106 (79%)	24 (73%)
ISS	25 (17–29)	25 (12–30)
Head AIS	5 (4–5)	5 (3–5)
Time: Injury to prehospital assessment (min)	31 (20–41)	31 (21–44)
Time: Injury to hospital (min)	83 (69–99)	88 (76–100)
GCS on scene	5 (3–7)	3 (3–5) **
First observation on scene
HR	82 (68–97)	103 (50–120)
SBP	142 (137–147)	73 (0–110) **
ED arrival observations
HR	86 (72–99)	104 (94–113) **
SBP	139 (117–159)	107 (94–119) **
Lactate	2.4 (1.4–3.3)	6.1 (1.7–10.9) **
Coagulopathy on arrival to hospital, n (%)
INR >1.2 OR PT >17.4	19 (15%)	13 (43%) **

AIS, Abbreviated Injury Scale; ISS, Injury Severity Score; GCS, Glasgow Coma Score; HR, heart rate (beats/ minute); SBP, systolic blood pressure (mmHg); INR, International normalized ratio; PT, prothrombin time. *Signifies P < 0.05 and **P < 0.01 comparing CVD with non-CVD groups (Mann–Whitney U or chi-squared tests).

Figure 2.

(A) Age distribution (%) In years. (B) Mechanism of injury. RTC: Road traffic collision. (C) Head AIS. Abbreviations: AIS: Abbreviated Injury Scale. * Signifies P < 0.05 and **signifies P < 0.01 when comparing CVD with non-CVD groups (Mann–Whitney U or chi-squared tests). (D) Pathology on CT head. EDH: Extradural hemorrhage; SDH: Subdural hemorrhage; SAH/IP/IV: subarachnoid hemorrhage/intraparenchymal blood/intraventricular blood; DAI: diffuse axonal injury; HIE: hypoxic ischemic encephalopathy. **signifies P < 0.01 when comparing CVD versus non-CVD groups (Mann–Whitney U or chi-squared tests).

Pre-hospital administration of HTS was the same between the groups, but pre-hospital and 24-hour blood component requirements were higher in the CVD patients (24-hour all blood products: 3 (0–8) vs. 0 (0–0) units, P < 0.001) (Table 2). In the HEMS 1 year cohort, there was no difference between the groups in the proportion of patients treated with a pre-hospital anaesthetic (CVD 90% vs non-CVD 94%), with a higher proportion of patients with CVD receiving Phenylephrine (CVD 53% vs non-CVD 6% P < 0.01).

Table 2.

Patient treatment and outcomes

Variable	Non-CVD (n 135)	CVD (n 33)
Pre-hospital treatment
Drug assisted intubation (%)	47 (94%)	17 (90%)
Hypertonic saline (NaCl 5%) n (%)	40 (30%)	13 (39%)
Blood products (PRBC)	0 (0)	0 (0–2) **
Blood products (FFP)	0 (0)	0 (0–1) **
24 hour blood transfusion (units)	0 (0)	3 (0–8) **
Neurosurgical intervention	48 (43%)	13 (43%)
ICU length of stay (days)	11 (3–19)	12 (2–25)
Vasopressor days	2 (0–5)	2 (0–4)
Total LOS (days)	28 (13–47)	32 (10–61)
Time to death (days)	2.5 (2–6)	2.5 (1–6)
28-day mortality	42 (31%)	20 (61%) **

ICU, intensive care unit; LOS, length of stay. *Signifies P < 0.05 and **P <0.01 comparing CVD with non-CVD groups (Mann–Whitney U or chi-squared tests). Pre-hospital Intubation was only recorded in the LAA cohort; hence, it was not documented in n = 85 (non-CVD) and n = 14 (CVD) patients. Patients who had an intubation without drugs will not have been included.

Overall 28-day mortality was 37%. Mortality was greater in CVD vs non-CVD iTBI (61% vs 31%, P = 0.002), however for patients with AIS 5 there was no difference (27% vs 25%, P = 0.262) (Fig. 3A, B). Length of stay in survivors and time to death were similar between groups (Table 2) and in survivors there was no difference in 28-day discharge destination (Fig. 3B). In CVD survivors, 38% (13/33) were discharged home, and 4/18 (22%) patients with recorded GOS had good neurological outcome (GOS 4–5).

Figure 3.

(A) Mortality associated with AIS. AIS: Abbreviated Injury Scale. (B) 28-day discharge destination. * Signifies P < 0.05 and **signifies P < 0.01 when comparing CVD versus non-CVD groups (Mann–Whitney U or chi-squared tests).

We performed univariant logistic regression examining five variables predicted to influence presence of CVD in the pre-hospital phase – age, lactate, sex, pre-hospital pupil reactivity, and pre-hospital GCS. Pupil reactivity (P = 0.001), pre-hospital GCS (P = 0.005) and lactate (P = 0.000) were significantly associated with CVD. On multivariant analysis using these five variables, only pre-hospital pupil reactivity (P < 0.05) and lactate (P < 0.05) were independently associated with CVD (Table 3).

Table 3.

Logistic regression analysis: dependent variable: CVD

	Univariate				Multivariate
Variable	Odds ratio	95% CI		P-value	Odds ratio	95% CI		P-value
		Lower	Upper			Lower	Upper
Age	0.997	0.977	1.017	0.764	1.036	0.983	1.091	0.190
Sex (male)	1.371	0.575	3.269	0.477	1.130	0.087	14.745	0.925
GCS pre-hospital	0.723	0.577	0.905	**0.005	1.135	0.667	1.931	0.642
Pupil reactivity pre-hospital	7.969	2.398	26.481	**0.001	7.971	1.183	53.736	*0.033
Lactate	1.375	1.196	1.581	**<0.001	1.382	1.039	1.836	*0.026

CVD, cardiovascular dysfunction; GCS, Glasgow Coma Score; CI, confidence interval. *Signifies P < 0.05 and **P < 0.01.

Discussion

This observational study has identified a critically injured cohort of severe isolated TBI patients of which one in five demonstrated physiological shock in the hyper-acute pre-hospital phase following injury. This cohort specifically excluded patients who may have exhibited CV dysfunction with a higher conscious level (GCS >8) or those with polytrauma, in which extra-cranial injuries may account for physiological CV changes. Despite this, the prevalence of CVD is higher than previous reports of hypotension for iTBI patients in-hospital (13%) or in pre-hospital care (12%).^[9,19] This study demonstrates that physiological shock in TBI occurs rapidly after injury and is frequently detectable within the first 30 minutes. This supports the theory that the critical phase of TBI occurs pre-hospital,^[7] and highlights the potential importance of accurate diagnosis and optimization of TBI management from the earliest point of clinician contact. Further work is required to better understand this clinically significant phenomenon and in turn determine whether any targeted early intervention would lead to improved patient outcomes.

TBI is a heterogeneous spectrum of pathology with AIS scoring used as a measure of severity, based on anatomical pathology identified clinically and radiologically, but is limited in determining the precise extent of neurological damage. We could not demonstrate an association between CVD and a specific type, or site of intracranial injury. Whilst Head AIS 5 represented the greatest proportion of CVD (42%) and non-CVD (53%) patients, it is interesting to note that patients with CVD had a higher proportion of mild TBI (AIS 3) than those without CVD, and there was no difference in mortality between groups with Head AIS 5. It is possible that CVD patients with lower AIS scores may present with features resulting from impact brain apnea and cardiac arrhythmias, some of which may be associated with good recovery^[12,23], and at the higher AIS scores features of cardiogenic shock may be apparent.^[23-25] Further research is needed to better understand the clinical features of CV dysfunction in this cohort.

The management of undifferentiated shocked trauma patients remains a challenge particularly in pre-hospital care. The recommendations for the treatment of inadequate cerebral perfusion pressure in confirmed TBI patients varies across international guidelines but includes the use of HTS, colloid boluses, inotropes, and vasopressors.^[26,27] In patients at risk of hemorrhagic shock with TBI one randomized control trial has found improved outcomes with early administration of plasma in the pre-hospital phase.^[28] In this study, patients with CV dysfunction received more blood components within 24 hours of injury despite having an isolated TBI, and we found significantly higher use of pre-hospital phenylephrine in the HEMS patients. These findings suggest that clinical interpretation of shock remains a challenge, particularly in the context of polytrauma where severe TBI may coexist with other significant injuries. Differentiating the bleeding patient from a severe iTBI with CV dysfunction, prior to definitive imaging, is not straightforward and will influence the treatments given and physiological treatment targets.

In order to identify the optimal treatment strategy for patients with TBI and CVD, it is necessary to understand the mechanisms underlying the CV features seen. Trauma induced secondary cardiac injury has been well described and associated with hyperacute elevations in inflammatory mediators^[29]. The pathophysiology of CVD following TBI remains less clear. Animal models have demonstrated an acute stress response (tachycardia and hypertension) to TBI, which has been attributed to a systemic catecholamine surge^[30,31]. Further, adult patients suffering TBI have demonstrated correlation between plasma catecholamine concentrations on arrival to hospital with ISS, GCS, functional outcome, and mortality^[13,14].

In addition to the effects of a systemic catecholamine surge on cardiac function (increased systemic vascular resistance, afterload and myocardial oxygen demand), there is evidence to support the impact of increased NA release directly to the heart. In vitro studies demonstrate myocyte hypercontraction and cell death following prolonged exposure to NA and support the hypothesis that excessive catecholamine release causes calcium overload and cAMP-mediated cardiac toxicity^[16]. Further, post mortem papers have demonstrated contraction band necrosis in the myocardium of patients who have an isolated TBI in the context of normal coronary arteries^[15].

It is likely that the mechanisms driving the clinical features described in this study are multi factorial and in addition to the systemic and local catecholamine effects, likely include: insular cortex and hypothalamus dysfunction affecting sympathetic nervous system activation and dysfunction;^[32] an acidotic hypercarbic environment following a period of apnea and dysventilation that may be unfavorable to a myocardium; and a systemic inflammatory mediator response.^[17] This serves to support the concept that TBI may provoke a multi-system response that results in the clinical presentation of CVD and shock. Significantly impaired cardiac output will result in inadequate cerebral perfusion and thereby oxygen delivery. This leads to secondary brain injury and ultimately cell death. As such, it is imperative that optimal management is initiated at the earliest opportunity. Further research is needed to elicit the primary driving mechanism for CV dysfunction and the best treatment strategy for the management of shock in the undifferentiated trauma patient at risk of severe TBI.

Multiple in-hospital studies have demonstrated poor outcome associated with cardiac dysfunction in TBI.^{[10,25,33-35]} This study demonstrates that CVD in the hyper-acute pre-hospital phase is associated with a significantly higher mortality (62% vs 33%). It has been suggested that the neurocardiac effects associated with TBI may be reversible if managed appropriately and early.^[18] Despite higher mortality, there remained a proportion of patients with CV dysfunction that were discharged home neurologically intact (4/18 (22%) in those who had GOS recorded at discharge). Further, there was no difference in the proportion of patients discharged to their usual residence, or in GOS at discharge, between the two groups. This implies there may be some hope for this group of severely injured patients who have historically been expected to have a poor prognosis.

Limitations

This study has several limitations including those inherent of an observational study. The cohort is restricted to patients with GCS <9 and of adult age. This study excluded children due to the variety in “normal” physiological variables for different ages but recognizes the need for further research to identify the prevalence of CV dysfunction following TBI in the pediatric cohort. It relies on accurate interpretation of GCS and documentation of physiology; both of which are subject to human error. CVD as determined by HR and blood pressure is pragmatic and clinically relevant but is just one way of determining evidence of dysfunction. Whilst we were able to analyses all observations for CVD prior to pre-hospital emergency anesthesia in the HEMS cohort, there was only a single initial pre-hospital observation to interpret with the ACIT cohort. As such, the prevalence of dysfunction may be greater than described in this study. Regarding the study cohort, we recognize it is possible that an acute medical event could conceivably precede a traumatic injury, and as such may play some part in the clinical presentation of the patient. However, the study cohort are younger than patients classically associated with such medical events (acute myocardial infarction or cerebrovascular event). It is also feasible that patients included in the study may have been taking medication that could impact their physiological response to trauma. We have used an internationally recognized ISS system to define a cohort of patients with severe isolated TBI, thus excluding any patients who have injuries in other body regions that are deemed to be significant. Regardless, it is possible that over time multiple lesser injuries could result in blood loss and shock if left untreated. If this were the case however, we anticipate this would be seen significantly later in the patient journey than the physiology being demonstrated by this cohort. To achieve a study cohort of meaningful size, it was necessary to combine two cohorts. The inclusion and exclusion criteria were the same with no crossover of patients between them. Despite this, the study sample size is relatively small. Primary outcome was set at, and limited by, 28-day mortality and GOS at discharge which was only available in the HEMS one-year cohort with a proportion unrecorded. However, the cohort identified is a pure sample of isolated severe TBI, both clinically represented by GCS <9, and anatomically according to AIS.

Conclusion

One in five patients with severe iTBI demonstrate early CV dysfunction. This occurs within a short timeframe from injury and does not correlate with AIS score or intra-cranial distribution of blood identified on CT. CV dysfunction is associated with increased overall mortality, coagulopathy and hypoxic ischemic encephalopathy. The mechanisms driving this phenomenon and the optimal management to this response remain poorly understood. However, the management of secondary insult and measures to correct physiological instability may be even more important in the pre-hospital phase of care as after arrival in hospital. Despite a higher mortality, neurological outcome is highly variable in those with CV dysfunction across the iTBI severity spectrum. Given the observational nature of the body of evidence to date, further research is needed to define the early pathophysiology of CV dysfunction in TBI and to determine whether early intervention is beneficial, to improve outcome for this subgroup of iTBI patients in the future.

Ethical approval

The ACIT-2 study was approved by a research ethics committee (REC reference: 07/Q0603/29). Retrospective analysis of deidentified data did not require further ethics approval. The LAA cohort met National Institute for Healthcare Research (NIHR, UK) approval for service evaluation and formal ethical approval was therefore not required. This service evaluation was approved by the Research & Clinical Effectiveness Lead at LAA and registered as a service evaluation at Barts NHS Trust, Kings College Hospital NHS Trust, Imperial NHS Trust and St George’s NHS Trust.

Consent

Not applicable.

Sources of funding

None.

Author’s contribution

F.B.: conceptualization (lead), methodology (equal), investigation (equal), writing – original draft (lead), writing – review & editing (equal), formal analysis (lead); MW, A.S. M.S., A.P., D.S.: investigation (equal), writing – review & editing; G.G.: writing- review & editing, project administration (lead); D.L.: writing- review & editing, supervision; R.D.: conceptualization (equal), methodology (equal), writing- review & editing (lead), supervision (lead).

Conflicts of interest disclosure

The authors declare that they have no competing interests.

Research registration unique identifying number (UIN)

ISRCTN12962642.

Guarantor

Flora Bird and Ross Davenport

Provenance and peer review

Not applicable.

Data availability statement

De-identified data from this study are stored in a safe repository in keeping with NHS Information Governance standards and are available on request. The datasets used and/or analyzed during this study are available from the corresponding author on reasonable request.

Presentation

The Big Sick Conference, Zermatt, 2024.