Patients with endocranial hypertension progress into brain death (BD) when intracranial pressure (ICP) exceeds the systolic arterial pressure, thus leading to cerebral circulatory arrest,1–5 being the intracerebral hemorrhage (ICH) the most common etiology in Spain followed by traumatic brain injury (TBI), and subarachnoid hemorrhage (SAH).3 Donors who die under conditions of brain death are still the main source of organs for transplants,6 but over the last few years there has been fewer patients dead due to neurological criteria,7 which is a universal problem since the number of organs available for transplant drops. In Spain, BD represents 2.3% of all in-hospital deaths, and 12.4% of all deaths reported in the intensive care unit (ICU) setting.8 In our setting, a European study confirms that 7.8% of the patients hospitalized at an ICU die under conditions of BD.9

We know that most patients with severe brain injuries who progress into BD do so within the first 72 h.3 However, being able to identify what patients will progress into BD and when they’ll do so is not an easy task. Brain injuries have a varied etiology, a different pathophysiological behavior, and great clinical complexity. Several studies have identified clinical and radiological signs10–15 predictive of progression into BD. However, in the clinical practice, there is no such thing as a tool to predict it with a high degree of certainty. Being able to identify, at admission, what factors are associated with early BD can help during the therapeutic management of patients, the decision-making process such as whether to proceed with life-sustaining treatment limitation (LSTL) or during the identification of potential organ donors.16

The objective of this study was to establish the existence of differences between clinical and radiological elements seen in patients who progress into early BD defined as death occurring within the first 24 h compared to those patients who died under conditions of BD after 24 h.

From 2015 through 2017, both years included, a total of 3834 patients were admitted to the ICU with a mean age of 60.3 ± 15.8 years, and an APACHE II score of 18.0 ± 7.9. The mean ICU stay was 7 ± 12.5 days. Out of all the hospitalized patients, 724 died (mortality rate, 18.8%), and 166 of these did so under conditions of BD (144 of them were organ and tissue donors). BD represented 22.9% of all deaths reported at the ICU setting, and 3.4% of all in-hospital deaths reported.

Radiological variables

Volume of hematoma: a total of 85 patients had a hematoma of supratentorial location with a mean volume of 86.5 mL, CI 0.95 [73.3–99.7], and 12 had an infratentorial hematoma with a mean volume of 23.8 mL, CI 0.95 [11.0–36.6]. No statistically significant differences were found regarding the volume of the supratentorial hematoma (90.9 mL in early BD vs 82.7 mL in late BD; P = .54) or the volume of the infratentorial hematoma (26.2 mL in early BD vs 23.0 mL in late BD: P = .83).

Extra-axial lesion: no patient with epidural hematoma was ever reported. In the series, 33 patients had subdural hematomas (16 from the in early BD group >24 h, and 17 from the BD group <24 h). The mean maximum thickness measured in mm in the first group was 9.59 compared to 16.23 in the second group; P = .053.

Brain midline shift: 10.7 mm in the early BD group vs 7.8 mm in the late BD group with a statistically significant difference (P = .045). For every increase of 1 mm in the brain midline shift, the risk of early BD goes up by 1.06 times (P = .047).

Optic nerve sheath: its diameter was measured between 3 mm and 10 mm away from the eyeball (Fig. 1). Results are shown on Table 1. No statistically significant differences were seen between the diameter of the optic nerve sheath and early BD.

Figure 1.

Measurement of optic nerve sheath on the cranial CT scan.

(0.26MB).

Table 1.

Diameter of the optic nerve sheath measured in mm.

CT scan at admission	BD ≥ 24 h	BD > 24 h	P
Left eye measured:
3 mm away from the eyeball	5,5	5,7	.23
10 mm away from the eyeball	4,4	4,4	.70
Right eye measured:
3 mm away from the eyeball	5,4	5,5	.60
10 mm away from the eyeball	4,3	4,3	.82
Mean:
3 mm	5,4	5,5	.60
10 mm	4,3	4,4	.86

Hydrocephaly: a total of 91 patients had ventriculomegaly, and 38 of these showed periependymal transudation. The presence of transudation kept a statistically significant correlation with progression into early BD (OR = 2.74, P = .021).

Total basal cistern effacement (36 patients from the early BD group vs 24 patients from the late BD group) (OR = 255 and P = .005), sulcal effacement (OR = 2.28; P = .013), the loss of cortico-subcortical differentiation (OR = 3.96; P = .0001), and the lack of suprasellar cistern (OR = 2.55; P = .005) kept a statistically significant correlation with progression into early BD. The partial basal cistern effacement was not associated with early BD (Table 2).

Table 2.

Signs of cerebral herniation and brain swelling on the cranial CT scan.

CT scan at admission	BD > 24 h	BD ≤ 24 h	P
	N = 85	N = 71
Signs of cerebral herniation
Basal partial cistern effacement	57	52	.4
Basal total cistern effacement	24	36	.005
Tumefaction/diffuse cerebral edema
Sulcal diffuse effacement	38	46	.013
Loss of cortico-subcortical differentiation	12	28	.0001
Ventricular collapse	15	15	.68
Absent suprasellar cistern	41	48	.015

When the multivariate analysis was performed progression into BD was associated with the loss of cortico-subcortical differentiation (OR = 2.81; P = .041), and GCS scores <5 (OR = 2.13; P = .05). The increase of 1 point in the GCS reduces the possibilities of progression into early BD by 1.4 times (P value = .013).

Neurological severity scores contribute to estimate the chances of death or functional prognosis, but without a high degree of certainty,17–20 which is why studies look for other factors that may be associated with progression into BD to optimize the clinical decision-making process at admission. Some authors have come up with the concept of «imminent BD» defined as patients with catastrophic cerebral injuries, Glasgow 3, lack of 3 or more trunk reflexes, and on mechanical ventilation.21,22 However, this entity is a highly evolved clinical situation that is not useful in the routine clinical practice. Humbertjean et al.12 describe 6 determinant factors to anticipate progression into BD: Glasgow scale scores ≤6 before sedation, volume of hematoma >65 mL, signs of herniation on the CT scan, hydrocephaly, early systolic arterial blood pressure >150 mmHg, and a past medical history of alcohol abuse with which they try to identify, within the first 24 h, patients with high chances of progressing into BD. Other studies find that the determinant factors of poor progression are old age, low scores on the Glasgow scale or high NIHSS scores, infratentorial location, arterial hypertension, anticoagulation and/or coagulopathy, ventricular hemorrhage, and hydrocephaly.13,23 In our study we did not find a statistically significant correlation with the patients’ past medical history or with their age, sex or even diagnosis. Consistent with the medical literature available, we did find that the Glasgow scale score at admission was a good predictor. In our case, the GCS score was 3 in 68.8% of the patients from the early BD group vs 38.2% of the patients from the late BD group (P = .0001) whereas GCS scores <5 were clearly associated with early BD. On the other hand, patients with intracranial hypertension treated with osmotic therapy, neuromuscular blockers, and decompressive surgery progressed into late BD as expected.

Regarding the radiological variables, traditionally we know that the presence of signs of endocranial hypertension on the CT scan is a severity indicator being the volume of the lesion one of the determinant factors of poor prognosis;14,24 regarding bleeding, several studies correlate the volume of hematoma with short- and long-term mortality.11,15,18,26–30 (Table 3). The presence of intraventricular blood and its volume, the infratentorial location of the hemorrhage or the brain midline shift are also determinant factors of poor prognosis.13,26,28,31,32 In patients with intracerebral bleeding, the lack of corneal reflex and the «whirlwind sign» seen on the CT scan (Fig. 2), indicative of active bleeding inside the hematoma, have been associated with progression into BD10 due to the greater chances of hematoma spread same as the presence of spot sign on the CCTA that is also indicative of active hemorrhage.33 In our series all patients died under situations of BD and showed signs of endocranial hypertension on the CT scan; the radiological signs that, in our setting, are significantly associated with progression into BD within the first 24 h are brain midline shift, the presence of active hydrocephaly, basal total cistern effacement, sulcal effacement, the loss of cortico-subcortical differentiation, and the lack of suprasellar cistern.

Figure 2.

Hemorrhage in right basal ganglia with ventricular spill and brain midline shift. The inside of the hematoma shows the «whirlwind sign», the arrows point at areas of different densities; the small hypodense areas correspond to foci of active bleeding through arterioles and venules.

(0.14MB).

The optic nerve sheath accumulates cerebrospinal fluid and increases its diameter in intracranial hypertension. Although there is no unanimity in the medical literature on the measuring site or the diameter required to consider endocranial hypertension, some authors find that diameters >5 mm to 6 mm are often followed by ICP >20 mmHg.34,35 We measured the optic nerve sheath between 3 mm and 10 mm away from the eyeball and found no significant correlation with an early progression into BD; in the measurements of the optic nerve sheath taken 3 mm away from the eyeball, the mean diameter found was > 5 mm in both groups of patients (early and late BD). Some of this study limitations are its retrospective nature, the fact that it only analyzed patients who progressed into BD, and that it was conducted in a single center only.

The scarcity of organs for transplants is a global problem that has led to the implementation of strategies to increase donation. The Non-therapeutic elective ventilation (NTEV) (in the Western world), and Donation-oriented intensive care (DOIC) (in Spain) initiatives promote the ICU admission of patients with cerebral catastrophic lesions without therapeutic options to promote donation.11,16,36 However, one of the difficulties of DOIC is to predict, with a certain degree of certainty, what patients will eventually progress into BD and when will they do so. In this sense, our study is interesting despite its limitations because it provides some of the most early prognostic factors thus facilitating the intensivist’s difficult time whenmaking a decision on an ICU admission when it comes to LSTL or patient maintenance as a potential organ donor under situations of BD.

Dolores Escudero designed the study and wrote the manuscript.

Iván Astola participated in the study design, statistical analysis, and the manuscript draft and proofread.

Ángela Meilán, and Débora Vizcaino collected the radiological data, analyzed the images from the cranial CT, and reviewed the manuscript.

Salvador Balboa, Blanca Leoz, Cecilia del Busto, Brígida Quindós, Lorena Forcelledo, Lorena Martín, Estefanía Salgado, and Lucía Viña performed the data mining, and proofread and reviewed the manuscript.

None whatsoever.

None reported.