Brain tissue partial pressure of oxygen predicts the outcome of severe traumatic brain injury under mild hypothermia treatment

Abstract

Objective

The aim of this study was to investigate the clinical significance and changes of brain tissue partial pressure of oxygen (P_btO₂) in the course of mild hypothermia treatment (MHT) for treating severe traumatic brain injury (sTBI).

Methods

There were 68 cases with sTBI undergoing MHT. P_btO₂, intracranial pressure (ICP), jugular venous oxygen saturation (S_jvO₂), and cerebral perfusion pressure (CPP) were continuously monitored, and clinical outcomes were evaluated using the Glasgow Outcome Scale score.

Results

Of 68 patients with sTBI, P_btO₂, S_jvO₂, and CPP were obviously increased, but decreased ICP level was observed throughout the MHT. P_btO₂ and ICP were negatively linearly correlated, while there was a positive linear correlation between P_btO₂ and S_jvO₂. Monitoring CPP and S_jvO₂ was performed under normal circumstances, and a large proportion of patients were detected with low P_btO₂. Decreased P_btO₂ was also found after MHT.

Conclusion

Continuous P_btO₂ monitoring could be introduced to evaluate the condition of regional cerebral oxygen metabolism, thereby guiding the clinical treatment and predicting the outcome.

Introduction

Monitoring the function and metabolism of the brain has proved to be a key factor in the treatment of patients with severe traumatic brain injury (sTBI). Currently, routine parameters for detecting the oxygen content and blood supply in brain tissue include intracranial pressure (ICP), cerebral perfusion pressure (CPP), transcranial Doppler ultrasound, and jugular venous oxygen saturation (S_jvO₂). However, the application of these monitoring parameters in clinical practice is quite limited.¹ It is well known that the metabolic enzyme activity in brain tissue is not equal and TBI may only induce local oxygenation dysfunction. Thus, the overall oxygenation of the brain could be within the normal range after TBI, which is unable to obtain an objective evaluation on the local oxygenation. In addition, local oxygen content monitoring has already become the focus in this research field all over the world. The detection of brain tissue partial pressure of oxygen (P_btO₂) level is getting more and more attention from clinicians. As a new monitoring method, P_btO₂ has the advantage of reflecting the local oxygenation of brain tissue over other parameters.²

Experimental and clinical studies have reported that mild hypothermia treatment (MHT) has a protective effect on the brain function and metabolism, which could reduce a series of adverse effects, prevent the occurrence of increased ICP, and minimize the consequent injury.^3–5 In this study, we combined the evaluation of MHT efficacy with oxygen monitoring in the management of patients with sTBI. The parameters including P_btO₂, S_jvO₂, ICP, and CPP were continuously detected, and their correlations were also evaluated in order to further investigate the effect of MHT on the oxygenation in the brain tissue. These results may help guide the selection of clinical strategy and predict the outcome for sTBI.

Methods

Patients

A total of 68 patients with sTBI who underwent MHT from August 2010 to September 2011 in our center were enrolled. Inclusion criteria were that the patients who were diagnosed with acute sTBI (Glasgow Coma Score ≤8 at admission), and the traumatic time was defined as the time since the accident was <3 hours. Patients with severe lung, abdominal, or other injuries, severe cardiac arrhythmia, coronary heart disease, and diabetes mellitus were excluded. Written informed consent was obtained from all the patients. This study was approved by the ethics committee of the Affiliated Hospital of Logistics University of People’s Armed Police Force.

Mild hypothermia treatment

During MHT, the room temperature was set at 17°C. The patients were asked to lie on the temperature-controlled blanket (Blanketrol II; CSZ, Cincinnati, OH, USA) for physical cooling, and the habination mixture of 100 mg chlorpromazine and 100 mg promethazine in 500 mL water or sedative drug of morphine sulfate in 100 mL water was administrated by intravenous injection to decrease the rectal temperature to 35°C within 6 hours and maintain at 33°C–35°C. When ICP became normal within 48 hours, the rewarming was started by controlling the temperature of the blanket to increase the rectal temperature from 36.5°C to 37.0°C for 12 hours. The breathing machine was applied if necessary. The heart rate, blood pressure, and muscle tone were monitored.

Brain function and metabolism monitoring

All the patients were monitored four times an hour for 72 hours. P_btO₂, S_jvO₂, ICP, and CPP at a total of 10,944 time points were recorded and analyzed. Intubation of left radial artery was performed. ICP and CPP were detected by Intellivue Mp30 system (Philip, Amsterdam, the Netherlands). A 0.5 mm testing probe was inserted into the frontal lobe at a depth of 27–36 mm to detect P_btO₂. The brain oxygen and temperature detector LICOR-II (Integra NeuroSciences Limited, Plainsboro, NJ, USA) and Nova-H Monitor (Nova Biomedical Corporation, Waltham, MA, USA) were introduced to continuously examine P_btO₂ and S_jvO₂ throughout the MHT for 7 days.

Statistical analysis

All the statistical analyses were conducted by SPSS software 13.0. All the continuous data were shown as mean ± SD. The categorical data were tested by chi-square test. P_btO₂, S_jvO₂, ICP, and CPP data at different time points were tested by repeated-measure data analysis of variance. Rank-sum test was used to test the outcome among groups. It was considered to be statistically significant, if P<0.05.

Results

Demographic and clinical characteristics

There were 39 male and 29 female patients with sTBI. The onset age was 37.8±11.9 years (ranging between 18 years and 72 years; Table 1). Glasgow Outcome Scale Score at admission was 5.3±1.2 (ranging between 3 and 8). Thirty-one patients underwent surgery, and the other 37 patients accepted conservative therapy. Traffic accident (n=31), beating (n=21), and falling injury (n=16) were the causes of sTBI. Epidural hematoma (n=5), subdural hematoma (n=9), brain parenchyma hematoma (n=4), cerebral contusion and laceration (n=10), diffusive axonal injury (n=8), multiple injuries (n=13), cranial basal bone fracture (n=14), and brain hernia (n=8) were detected by craniocerebral computed tomography (CT) scan.

Table 1.

Demographic and clinical characteristics

	sTBI patients
Sex, n (%)
Male	39 (57.4)
Female	29 (42.6)
GCS score, mean ± SD (range)	5.3±1.2 (3–8)
Causes of sTBI, n (%)
Traffic accident	31 (45.6)
Beating injury	21 (30.9)
Falling injury	16 (23.5)
Treatment, n (%)
Surgery	31 (45.6)
Conservative therapy	37 (54.4)
Age (years), mean ± SD (range)	37.8±11.9 (18–72)
CT imaging, n (%)
Epidural hematoma	5 (7.3)
Subdural hematoma	9 (13.2)
Brain parenchyma hematoma	4 (5.9)
Cerebral contusion and laceration	10 (14.7)
Diffusive axonal injury	8 (11.8)
Multiple injuries	13 (19.1)
Cranial basal bone fracture	14 (20.6)
Brain hernia	8 (11.8)

Abbreviations: CT, computed tomography; GCS, Glasgow Coma Scale; sTBI, severe traumatic brain injury; SD, standard deviation.

Correlation of P_btO₂ with S_jvO₂, ICP, and CPP

The mean P_btO₂, S_jvO₂, and CPP were obviously increased, but ICP was decreased at 12 hours under MHT (Figure 1). At 24 hours and 36 hours, no remarkable changes of P_btO₂ and S_jvO₂ were observed. Further analysis showed that P_btO₂ and ICP were negatively correlated (r=−0.936, P<0.05). P_btO₂ and S_jvO₂ (r=0.966, P<0.05), and P_btO₂ and CPP (r=0.956, P<0.05) were positively linearly correlated after treatment with MHT for 24 hours. P_btO₂ <10 mmHg for >10 minutes was found in 56 patients, while only 21 patients had S_jvO₂ <55% for >10 minutes (Table 2). There were still a high percentage of patients who had normal CPP but low P_btO₂.

Figure 1.

Comparison of data on P_btO₂, S_jvO₂, ICP, and CPP at 12 hours, 24 hours, and 36 hours after MHT.

Notes: *vs 12 hours (P<0.05); ^#vs 24 hours (P<0.05).

Abbreviations: CPP, cerebral perfusion pressure; ICP, intracranial pressure; MHT, mild hypothermia treatment; P_btO₂, brain tissue partial pressure of oxygen; S_jvO₂, jugular venous oxygen saturation.

Table 2.

Detection of P_btO₂ and S_jvO₂ in 68 patients after MHT

Parameters	≥Mean value	<Mean value
SjvO2, n (%)	47 (69.1)	21 (30.9)
PbtO2, n (%)	22 (32.4)	46 (67.6)

Abbreviations: MHT, mild hypothermia treatment; P_btO₂, brain tissue partial pressure of oxygen; S_jvO₂, jugular venous oxygen saturation.

P_btO₂ as a prognostic factor for sTBI

All the patients were divided into two groups, based on P_btO₂ at admission. There were 46 patients in group A (P_btO₂ ≤10 mmHg) and 22 in group B (P_btO₂ >10 mmHg). Survival data showed that eight patients died and five were in vegetative state in group A; no patients died or were in vegetative state in group B (Table 3). Statistical analysis indicated that low P_btO₂ was significantly associated with poor outcome (P<0.01).

Table 3.

Association of P_btO₂ with patients’ outcome

Outcome	PbtO2		P-value
	≤10 mmHg (n=46)	>10 mmHg (n=22)
Dead, n (%)	8 (17.4)	0 (0.0)	0.145
Alive, n (%)
Vegetative state	5 (10.9)	0 (0.0)
Severe disability	4 (8.7)	6 (27.3)
Moderate disability	7 (15.2)	6 (27.3)
Good	22 (47.8)	10 (45.5)

Abbreviation: P_btO₂, brain tissue partial pressure of oxygen.

Discussion

In recent years, the application of P_btO₂ monitoring in the management of severe brain trauma is getting more and more attention. Compared with S_jvO₂, P_btO₂ has the advantage of being objective and rarely influenced.^6,7 Although P_btO₂ is a reliable parameter and could be monitored for a long observation period, it is still unable to manifest the overall brain function, especially in the probe placing site along with high metabolic area or low blood supply. Local P_btO₂ value may be misinterpreted as the P_btO₂ of the whole brain. Thus, there is controversy on the accuracy of P_btO₂. Gopinath et al⁸ demonstrated that patients without decreased S_jvO₂ had good prognosis. However, mortality in these patients was still much higher than that in patients without P_btO₂ <10 mmHg or with P_btO₂ <10 mmHg for <30 minutes. P_btO₂ may be decreased while S_jvO₂ is still within normal range, indicating that P_btO₂ as a parameter is more sensitive than S_jvO₂. In our patient cohort, the percentage of patients with low P_btO₂ was high. There were 56 patients with P_btO₂ <10 mmHg for >10 minutes, while 21 patients had S_jvO₂ <55% for >10 minutes.

We also found that P_btO₂ was positively correlated with CPP. P_btO₂ is mainly influenced by CPP, when arterial pressure of carbon dioxide and oxygen keeps stable. The increase in CPP could improve P_btO₂, which contradicts with the conclusion by Kiening et al.⁹ They demonstrated that P_btO₂ was stable when CPP was >60 mmHg, which was associated with the detection methods. In this study, we placed the probe in the ischemic area near the brain trauma, while Kiening put the probe in the non-ischemic area. In addition, automatic regulation dysfunction is always present in ischemic tissue. What is more, low P_btO₂ could coexist with normal CPP. However, normal CPP has no relation with good cerebral blood supply in patients with TBI.¹⁰ In acute brain trauma, both low CPP and changes of local cerebral vessels including microcirculatory injury and spasm, stenosis, and embolism of large vessels contribute to the ischemia.^11,12 Normal CPP could exclude the diagnosis of local ischemia and hypoxia in brain tissue.

ICP is another key factor in the treatment of brain trauma. Increased ICP directly decreases CPP, leading to the insufficient cerebral blood supply and low P_btO₂. McKinley et al¹³ have ever established a dog model of high ICP, and the similar results were found. Besides, hypoxia is associated with poor outcome in patients with brain trauma. van Santbrink et al¹⁴ analyzed 22 patients with TBI and proved that most patients had a low cerebral oxygen period after TBI. In our study, 46 patients underwent P_btO₂ <10 mmHg within 24 hours after TBI, eight of whom died and five patients were in vegetative state for 6 months. On the contrary, no death and vegetative state were observed in 22 patients with normal P_btO₂. All the evidence supports that P_btO₂ is closely correlated with patients’ outcome.

Conclusion

Our findings demonstrated that the combination of P_btO₂, S_jvO₂, ICP, and CPP could help guide the management and predict the outcome for patients with sTBI.