Intracerebral hypoglycemia and its clinical relevance as a prognostic indicator in severe traumatic brain injury: A cerebral microdialysis study from India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.177617
Source of Support: None, Conflict of Interest: None
Context: Traumatic brain injury (TBI) remains a major cause of morbidity and mortality worldwide. Largely, the prognosis is dependent on the nonmodifiable factors such as severity of the initial injury, Glasgow coma scale score, pupillary response, age, and presence of additional physiological derangements such as hypoxia or hypotension. However, secondary insults continue to take place after the initial injury and resuscitation. The study hypothesis in the present research article was that hypoglycemia is an independent outcome prognosticator in severe traumatic brain injury. The study aimed to assess the role of glucose monitoring in the brain parenchyma as an independent outcome prognosticator and also to study its association with plasma glucose levels.
Keywords: Decompressive craniectomy; glucose; microdialysis; traumatic brain injury
Traumatic brain injury (TBI) remains a major cause of worldwide morbidity and mortality. The prognosis is largely dependent on the nonmodifiable factors such as the severity of the initial injury, the Glasgow coma score, the pupillary response, the age of the patient, and the presence of additional physiological derangements such as hypoxia or hypotension., However, secondary insults continue to take place after the initial injury and resuscitation. With the advent of microdialysis (MD), the monitoring of the penumbric zone and of biochemical changes taking place inside brain tissue has become possible for the neurosurgeon. This has palyed the important role of transporting research in this arena from bench-side to bedside. The present study aimed to assess the role of glucose monitoring in the brain parenchyma and its association with plasma glucose for outcome prediction in patients suffering from severe traumatic brain injury.
The study included all patients of severe TBI (Glasgow Coma Scale [GCS] score ≤8) with a surgically treatable lesion. The patients included were >18 years in age. As part of the routine protocol, the patients underwent a noncontrast computed tomography (NCCT) of the head at admission along with an assessment of other systemic injuries. Patients who were pregnant, those with GCS score of 3 with fixed dilated pupils, or those who were hemodynamically unstable were not enrolled in this prospective, nonrandomized study. Informed consent and Institute review board permission was obtained prior to the inclusion of the patients in this study.
All severe traumatic brain injured patients with a surgically treatable lesion and/or raised intracranial pressure (ICP)/refractory intracranial hypertension (subdural hematoma, contusion, diffuse cerebral edema with refractory intracranial hypertension, that is, ICP >20 mm Hg for over 30 min) were subjected to decompressive craniectomy with augmentation duroplasty. Postoperatively, they were managed in the neuro-intensive care unit. At the time of dural closure, a single MD catheter, CMA-70 (a 20 kDa catheter), was inserted in the peri-contusional brain parenchyma at a depth of 20 mm and connected to the cerebral MD (CMD) pump which was preloaded with central nervous system (CNS) perfusion fluid. Few patients had catheters placed bilaterally. All patients were managed as per the standard treatment guidelines given by the Brain Trauma Foundation (BTF 2007). All patients also underwent a concurrent ICP monitoring using an intraparenchymal ICP catheter based monitoring. The patients also had invasive arterial pressure monitoring. Postoperative NCCT head was routinely done to confirm the position of the catheter(s) in the penumbric/intraparenchymal zone.
MD catheter probes (molecular weight cut-off of 20 kDa, 10-mm membrane, CMA-70, CMA MD) were placed in the peri-lesional tissue at the time of dural closure. The probes were perfused with CNS perfusion fluid at 0.3 µl/min (P000151, CMA MD) and samples collected for intracerebral glucose every hour. Concentrations of glucose, lactate, pyruvate, glutamate, and glycerol in the microdialysate were analyzed using the CMA-ISCUS flex MD analyzer [Figure 1].
The MD catheter was kept in place for a minimum of 3 days and up to a maximum of 5 days. Hourly data points were recorded for each individual. A data point was defined as the MD values acquired after each hour in addition to the neurophysiological parameters from the multimodal brain monitoring including the ICP, cerebral perfusion pressure (CPP), and mean arterial pressure.
Simultaneous analysis of plasma glucose was done on an hourly basis through an arterial line and analyzed using an arterial blood gas (ABG) analyzer. Plasma glucose levels were maintained between 70 and 140 mg% as per our neuro-intensive care protocol and hyperglycemia (serum glucose >200 mg%), if any, was managed by the insulin sliding scale. As this was a preliminary, noninterventional, observational study to study the correlation between cerebral glucose and serum glucose in severe TBI patients, no attempt was made to correct the glucose levels on the basis of CMD glucose levels in any case.
The normal extracellular glucose concentration in the human brain has not been well established. Normative data are dependent on several technical factors, including the perfusion rate. Careful assessment of the perfusion rate is required when comparing various studies. With the reduction in the perfusion rate to 0.3 uL/min, the in vivo recovery increases. In keeping with the international standards, the standard CMD based biochemical values for reference were kept as: Glucose ≥2 mmol/L (≥36 mg%); lactate <2 mmol/L; pyruvate 0.12 mmol/L; glycerol 20–50 micromol/L; glutamate 10 micromol/L; and, lactate/pyruvate ratio 15–20.
All data were tabulated in MS Excel 2011 and analyzed using IBM SPSS version 21 (IBM, USA). To calculate the correlation between plasma and MD glucose, the Pearson's correlation coefficient was used with a one-tailed test of significance. Student's t-test was used to calculate the difference in means between two groups. Significance was assumed at P ≤ 0.05.
The institutional ethical board permission was obtained for the study. All patients underwent decompressive craniectomy as per the standard management protocol. Informed and written consent for MD catheter insertion and monitoring was taken from a relative at the time of surgery[Table 1].
The mean age of the patients enrolled in the study was 31.76 (standard deviation [SD]: 10.71) years (range: 18–64 years). There were 21 males and four females in this group. The mean GCS before intervention was 5.36 (range: 4–8). The mean plasma glucose at the start of MD was 7.67 mmol/l with a SD of 1.55 (range: 5.23–8.44). The mean MD glucose at the start of the MD was 1.77 mmol/l with a SD of 1.59 (range: 0.1–7.7) [Table 2].
A total of 2116 corresponding arterial blood and brain MD readings were obtained. In our study, the mean random blood sugar (RBS) for the whole cohort was 7.39 mmol/l. In the good outcome group, 1303 values were obtained with a mean RBS of 7.264 mmol/l and a mean MD glucose of 1.841 mmol/l. In the bad outcome group, the mean RBS was 7.61 mmol/L with a mean MD glucose of 1.95 mmol/L. However, there was no significant difference in the mean RBS (P = 0.166) and mean cerebral glucose (P = 0.221) values amongst the good or bad outcome groups [Table 3].
There was a poor correlation between the plasma and cerebral glucose for the entire cohort (Pearson's correlation 0.01 with P = 0.642). No correlation was seen in the pooled glucose values in either the good Glasgow Outcome Scale (GOS) or the bad GOS groups. This suggests that brain MD glucose values are not reflective of changes in the systemic glucose values. Taking matched blood and MD glucose values individually for each patient, only 7 of the 25 patients (28%) had a significant correlation between plasma and cerebral glucose. Out of these, four patients belonged to the good outcome group, two to the bad outcome group, and one had the catheter placement in an uninjured lobe in the bad outcome group.
There was no significant difference in the mean RBS (P = 0.166) and the mean cerebral glucose (P = 0.221) values amongst good or bad outcome groups.
There was a significant difference in the incidence of hyperglycemia (RBS >10 mmol/L) between the two groups (P < 0.0001). However, the MD glucose values during the episodes of hyperglycemia did not show a significant difference between the two groups (P = 0.859).
There was a significant difference in the incidence of hypoglycemia (RBS <5 mmol/L) between the two groups (P = 0.0026). However, the MD glucose values during the episodes of hypoglycemia did not show a significant difference between the two groups (P = 0.455).
The plasma glucose values were divided into 6 clusters to assess the role of plasma glucose on cerebral glucose [Table 4] and [Figure 2].
Patients were also assessed with respect to the incidence of hypoglycemia and the low levels of MD glucose. A large proportion of MD glucose values below 2 mM were observed in both the groups, whereas low blood sugar values were seen in a higher proportion in the poor GOS patients. The difference between the two groups while comparing episodes of MD hypoglycemia during the hypoglycemic episodes (RBS <5 mmol/L) was significant (P = 0.0026). Authors opine that this reflects an increased susceptibility of brain tissue to hypoglycemia in the poorer outcome group [Table 5].
Fifteen of the 25 patients enrolled in the study had a good GOS (4, 5) at 3 months. The good outcome group had fewer episodes of brain hypoglycemia during systemic hypoglycemia (P = 0.0026). Neither the mean blood glucose values nor the mean cerebral glucose values predicted the outcome at 3 months. While comparing the daily mean values in both the groups, it was seen that MD glucose values showed a rising trend in the good outcome group. Similarly, the good outcome group also showed rising MD glucose to blood glucose ratios. Authors hypothesize that this indicates a recovering metabolic state in the brain tissue. Rising MD glucose levels or recovering the value of brain glycemic control can be taken as a real time, early bedside sign of a good neurological outcome [Figure 3].
Of the 15 cases with a good outcome (GOS 4 and 5), 4 cases showed a positive correlation (Pearson's correlation coefficient analysis) between the CPP and MD glucose values while 2 patients showed a negative correlation. Of the 10 cases with a poor outcome (GOS 1–3), only one case had a positive correlation between the CPP and MD glucose values.
It is well known that hyperglycemia and hypoglycemia need to be avoided to prevent increase in the underlying brain damage.,, Hyperglycemia exacerbates tissue acidosis and oxidative stress aggravating underlying brain damage which, in turn, promote the development of multiorgan failure., Hypoglycemia impairs energy supply causing metabolic perturbations and inducing spreading cortical depolarization. The NICE-SUGAR trial showed a significant increase in the mortality in patients subjected to the tight blood glucose range of 4.5–6.0 mmol/l compared with the conventional glucose control group with a blood glucose target of 10 mmol/l or less. In our study, patients of both the groups had their blood glucose maintained at an average of 7.39 mmol/l. Neutralizing dose of insulin was giving in the case of high blood sugar values. However, both hypoglycemic and hyperglycemic episodes were more common in the poor outcome group.
It was earlier thought that cerebral extracellular glucose concentration levels would change in parallel to the blood glucose values. This may, however, be an over simplification. Vespa et al., have shown that following head injury, there may be an increased utilization of cerebral glucose leading to lower MD glucose values. Schlenk et al., also found a poor correlation between the blood and extracellular blood glucose in a patient with subarachnoid hemorrhage. This may also be the case with TBI patients. Abate et al., attributed the poor control of MD glucose (based on blood glucose values) to the metabolic heterogeneity that may be expected in TBI patients. In our study too, a poor correlation was noted between the MD glucose and blood glucose values.
Vespa et al., also showed that MD glucose values were better predictors of outcome at 6 months as compared to blood glucose values. However, our study showed no statistical difference in MD values amongst the two groups at 3 months post-injury.
After decompressive craniectomy in severe TBI, there was a poor correlation between the plasma and MD glucose concentration. A high degree of variation was seen in the correlations for individual patients. Neither the mean blood glucose values nor the mean cerebral glucose values predicted the outcome at 3 months. The good outcome group had fewer episodes of both hyperglycemia and hypoglycemia.
The authors wish to thank staff nurse Ms. Jyoti Sohal (RN) at our institute for handling the CMD machine in the neurosurgery intensive care unit of our hospital.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]