Correlation of Cerebral and Subcutaneous Glycerol in Severe Traumatic Brain Injury and Association with Tissue Damage

Traumatic brain injury (TBI) can be described as mild, moderate, or severe (sTBI) and is often classified on the basis of level of consciousness and/or the length of posttraumatic amnesia. sTBI is the main cause of morbidity and mortality among young adults younger than 40 years of age, and the mortality rate is estimated to be 10 per 100,000 inhabitants per year in Sweden [1]. The overall levels of in-patient care of TBI have seemed to decrease, but there is an increase in older adults [2]. A similar trend is reported in other European countries [3]. In a recent review, the global burden, consequences, and health economy costs were found to be very large [4].

Glycerol is a molecular component of triglycerides stored in the human fatty tissue. Triglycerides are decomposed into free glycerol and fatty acids through sympathetic activity. Glycerol does not normally pass the blood–brain barrier (BBB). In the central nervous system, glycerol is only present as phospholipid in cell membranes. Therefore, glycerol quantified intracerebrally is used as a measure of brain cell damage and death [5]. This damage may be caused, for example, by direct mechanical damage and metabolic distress, including inflammation. Because the BBB may be damaged by the injury, it is important to reduce further damage to the BBB and to avoid treatment measures that may worsen the leakage through the BBB [6, 7]. In case of a disrupted BBB, there may be an increased transfer into the brain of blood-borne glycerol released from extracranial sources, such as adipose tissue. There is some evidence that peripheral glycerol may penetrate into the central nervous system after brain injury [8]. If this occurs, the value of glycerol as a marker of cerebral tissue injury could be questioned.

Osmotherapy is one of the most commonly used emergency treatment measures in patients with sTBI with high intracranial pressure (ICP). Intraarterial administration of mannitol may open the BBB [9]. Whether this is true for intravenous administration of mannitol, in doses used in the clinical management, leading to influx of peripheral constituents in the brain, is not clear. If osmotherapy would further open an already damaged BBB, allowing blood constituents to penetrate the brain, measurement of, for example, brain extracellular glycerol, would be hard to use as a marker of brain damage. Surprisingly, there are almost no publications reporting on any correlation between cerebral and other sources of glycerol measured temporally.

Brain tissue damage can be quantified on the basis of levels of protein S-100B and neuron-specific enolase (NSE) in the peripheral blood. S-100B is a calcium-binding protein that is mainly found in the glial cells in the central and peripheral nervous system. Measured in blood, it has an estimated sensitivity of 96% and a specificity of 31% to detect intracranial injuries seen on computed tomography (CT) scans [10]. NSE is a glycolytic enzyme found in the cytoplasm of neurons but also in erythrocytes and platelets. It has shown a varying ability to detect brain injuries seen on CT scans [10, 11].

Increased ICP has been shown to be a prognostic factor for poor outcome after TBI [12]. At the university hospital in Umeå, a modified Lund concept is used in the treatment of TBI. The guidelines used in the Lund concept are focused on volume regulation of the brain and microcirculation and aim at controlling ICP and optimizing microcirculation in the penumbra zone. The goal is to reduce or prevent secondary injury, such as ischemia and edema [13, 14].

We hypothesized that cerebral glycerol (glycerol_brain) levels would not be associated with peripheral subcutaneous adipose glycerol (glycerol_sc) levels. If a correlation was observed, this might call into question the use of brain glycerol as a solid marker of brain tissue damage not mirrored in the brain CT classification, biochemical biomarkers, and clinical outcome. The first aim was to study the relationship between intra- and extracerebral glycerol (glycerol_brain/sc) levels. The second aim was to study any associations between glycerol_brain/sc and measurements of brain tissue damage, overall injury level, injury type, and clinical outcome.

Demographics and general data are shown in Table 1. All patients had sTBI and a GCS ≤ 8 and were intubated and in need of neurointensive care. The distances from microdialysis catheter tip to lesion are presented in Table 1. Types of accidents were traffic accidents, falls, sports, and assaults (see Table 2). No significant difference between the groups was observed.

There were 25 individuals with a positive glycerol_brain/sc correlation (individual Pearson’s R² coefficients between 0.0002 and 0.4144) and 17 with a negative correlation (individual Pearson’s R² coefficients between 0.000001 and 0.4610). The corresponding regression plot results were significantly different (Table 3). Figure 2 summarize these results. Prostacyclin had no significant effect on the association of glycerol_brain/sc in either group (t-test). The mean value of glycerol_brain in the first 96 h was slightly higher in the group with a positive correlation of glycerol_brain/sc compared with the group with a negative correlation (Table 3). These results are summarized in Fig. 3. There was no significant difference in glycerol_sc levels between groups with positive or negative correlation of glycerol_brain/sc.

There was no significant difference in mean age (t-test) or GCS (Wilcoxon’s rank-sum test) in the two groups (see Table 1).

No significant association between initial CT Rotterdam score or that after 24 h and glycerol_brain/sc (Wilcoxon’s rank-sum test) (Table 1) was observed. There was a positive correlation between the Rotterdam score at 24 h after injury (but not the initial score) and mean glycerol levels in the brain (ρ = 0.3565, p = 0.0221, Spearman). No such correlation was observed with the peripheral glycerol levels. There was a significant association of S-100B_max, S-100B_AUC, NSE_max, and glycerol_brain/sc (Wilcoxon’s rank-sum test) (Table 3). We found no association between other brain metabolic markers (glucose, lactate, lactate/pyruvate) and the glycerol_brain/sc patterns (Table 3).

MAP (mean 96 h) did not differ between the two groups. No association between ICP or CPP levels (mean 96 h) and glycerol correlation was observed (Table 1).

There was no difference in the proportion of decompressive hemicraniectomies between the groups (χ² test).

Type of injury, ISS, and APACHE II had no impact on the results, and there were no significant differences in these parameters between the groups (Table 1), nor was there any correlation between the brain and peripheral glycerol levels or between the ratio of brain to peripheral glycerol and ISS and APACHE II.

At the 6-month follow-up, the median GOSE was 4 (1–8) in the group with a negative glycerol_brain/sc correlation and 5 (1–7) in the group with a positive correlation (ns), Wilcoxon’s rank-sum test).

We have shown that in 60% of patients with sTBI, peripheral glycerol may flux into the brain. This effect is associated with worse brain tissue damage, as measured by S-100B and NSE.

Glycerol has been considered as a marker for brain cell damage after TBI [5] and may be correlated with clinical outcome. This damage may be a direct mechanical damage or damage due to metabolic distress. In a situation with a flow of glycerol from the blood into the brain due to, for example, a BBB disruption, the intracerebral glycerol level ceases to mirror the intracerebral condition. In such a situation the glycerol levels in the brain may be misinterpreted to indicate an increased breakdown of brain tissue. In 60% of the patients, a positive correlation was observed. This may be interpreted as there being a possibility of an inflow of glycerol to the brain and may be due to, for example, some kind of BBB dysfunction. Therefore, one might suggest that some caution should be used in interpreting brain interstitial levels of glycerol as a marker of brain cell damage. The pathophysiological mechanism behind the negative correlation between extra- and intracerebral glycerol levels is interesting. One reason for this pattern could be that the peripheral glycerol level mirrors extracranial injuries, resulting in an altered glycerol ratio between peripheral and intracranial tissue. However, we could not identify any difference in type of accident, extracranial injuries, or level of injury severity, as measured by using ISS and APACHE II, nor was there a difference in peripheral glycerol levels between the groups. In addition, we found no indication of a difference in the metabolic patterns in the brain, as reflected in the interstitial concentrations of glucose, lactate, and lactate/pyruvate.

All patients in this study were treated with β-blockers (metoprolol) and α₂-stimulation (clonidine) to decrease the sympathetic nervous system drive, thus decreasing the general stress level and by extension, the inflammation process. Because subcutaneous glycerol release is promoted by an activation of the sympathetic nervous system, and we were actively trying to decrease this response, it is unlikely that the general stress levels of the patients influenced the results. We did not measure the blood level of glycerol, but it has previously been reported that subcutaneous glycerol levels measured by microdialysis follows the pattern of glycerol in the blood [20].

There was no significant association of ICP or CPP and glycerol_brain/sc. This can simply reflect effects of the treatment striving to normalize the ICP and CPP. It is more remarkable that the tissue damage categorized by using the Rotterdam score was not associated with extra- to intracranial glycerol correlations. However, surrogate measures of brain tissue damage, i.e., S-100B and NSE, showed more impairment in the patients with a positive correlation of glycerol_brain/sc. This may indicate that the greater the brain tissue damage, the greater the leakage of glycerol from extracerebral tissues (subcutaneous fat) into the brain, which may be due to a disrupted BBB.

We have previously shown that prostacyclin treatment is associated with a stabilization of the pressure reactivity of cerebral vessels [21] and the relation between cerebral glycerol and pressure reactivity in sTBI [22]. Interestingly, in the present study, prostacyclin had no clear-cut influence on the intra- to extracerebral glycerol correlation. Thus, one could speculate, that the previously reported effect of prostacyclin on pressure reactivity and glycerol is not influenced by an extra- to intracerebral glycerol flux.

Some support for our findings may be found in the publication by Meyer et al. [8]. They showed, in patients with cerebral infarction, that intravenous infusion of glycerol entered the brain and cerebrospinal fluid. In addition, glycerol given for treatment of high ICP in patients with brain edema after a large middle cerebral artery infarction resulted in a marked accumulation of brain interstitial glycerol, as measured through microdialysis [23]. To the best of our knowledge, no previous studies have been reported on de novo intra- to extracranial glycerol correlation in sTBI. Thus, it is impossible to compare our results with others. Because glycerol is clinically used as a very potent hyperosmotic drug in patients with life-threatening intracerebral processes, it is wise to consider that a disrupted BBB may cause leakage of glycerol into the brain tissue. One could speculate that hyperosmotic agents clinically used in attempts to reduce brain edema in patients with sTBI may further increase such leakage, leading to an increased osmotic tissue pressure. This may result in an aggravated tissue edema, a phenomenon that can be observed in these severely ill patients.

As with all clinical and experimental studies, there are strengths and weaknesses. The strengths of our study are that all patients had been treated and monitored in accordance with a predefined schedule. All patients were from a specified region and havd sTBI. Weaknesses include that we only studied patients with sTBI and therefore our results cannot directly be applied on patients with mild and moderate head injury. In addition, whether our results are applicable on patients treated by different treatment algorithms is unknown. Our study is too small to draw any certain conclusions. Therefore, further studies must validate or reject the present results.

There may be an influx of glycerol from extracerebral sources into the brain, confusing the interpretation of the cerebral glycerol levels. This eventual influx was paralleled with higher S-100B and NSE levels, indicating a worse brain tissue injury. We remind the clinicians and researchers that a damaged BBB, as seen in sTBI, may alter the levels of various substances, including glycerol in the brain. Awareness of this is important in the interpretation of the bedside data as well as in research.