High Arterial Glucose is Associated with Poor Pressure Autoregulation, High Cerebral Lactate/Pyruvate Ratio and Poor Outcome Following Traumatic Brain Injury

Patients with severe TBI were admitted to the Department of Neurosurgery at the University Hospital in Uppsala, Sweden. Out of 1001 TBI patients (2008–2018), there were 120 patients that met the study inclusion criteria, i.e., neurointensive care (NIC) with intubation, mechanical ventilation and monitoring of arterial blood pressure, ICP, and cerebral energy metabolism (cerebral microdialysis). Patients were treated in accordance with our standardized ICP-oriented treatment protocol to avoid secondary insults [14]. Treatment goals were ICP ≤ 20 mm Hg, CPP ≥ 60 mm Hg, systolic blood pressure > 100 mm Hg, central venous pressure 0–5 mm Hg, pO₂ > 12 kPa, arterial glucose 5–10 mmol/L (mM), electrolytes within normal ranges, normovolemia, and body temperature < 38 °C. The arterial glucose was monitored with repeated arterial blood gases. Hyperglycemia was corrected with short-acting insulin injections and in some cases intravenous insulin infusions for patients with diabetes mellitus.

ICP was monitored with either an intraparenchymal (Codman ICP Micro-Sensor, Codman & Shurtleff, Raynham, MA) or an intraventricular catheter drainage system (HanniSet, Xtrans, Smith Medical GmbH, Grasbrunn, Germany). Arterial blood pressure was measured invasively in the radial artery. Arterial blood gas (ABG) data were generally analyzed from the radial arterial line every fourth hour, more often if needed. PRx, traditionally calculated as the 5 min correlation of 10 s averages of ICP and MAP, was used in combination with a bandpass filter, limiting the analysis to oscillations with periods of 15–55 s (PRx55-15). It has previously been shown that PRx55-15 has better signal-to-noise ratio and correlation with outcome than the traditional PRx [9].

Cerebral energy metabolism was monitored with the 71 high cut-off microdialysis analyzer (MD) catheter with a membrane length of 10 mm and a membrane cut-off of 100 kDa (M Dialysis AB, Stockholm, Sweden). The catheters were perfused by means of a microinjection pump (106 MD Pump, M Dialysis AB) at a rate of 0.3 µL/min with custom-made sterile artificial cerebrospinal fluid (CSF) containing—NaCl 147 mmol/L (mM), KCl 2.7 mM, CaCl₂ 1.2 mM, MgCl₂ 0.85 mM, and 1.5% human albumin. Cerebral extracellular glucose, lactate, pyruvate, and urea were measured hourly, using a CMA 600 analyzer or the ISCUSflex Microdialysis Analyzer (M Dialysis AB). The MD urea was monitored to validate catheter performance [15]. The MD was placed in normal-appearing brain tissue in the right frontal lobe, adjacent to the ICP monitor.

Clinical outcome was assessed at 6-month post-injury, by specially trained personnel with structured telephone interviews, using the Glasgow Outcome Scale-Extended (GOS-E), containing eight categories of global outcome, from death to upper good recovery [16‐18].

Mean daily (24 h) values were calculated for each patient the first 3 days and for the entire 3-day-period (72 h) post-injury including ICP (mm Hg), CPP (mm Hg), PRx55-15, arterial glucose (mM), cerebral glucose (mM), cerebral lactate (mM), cerebral pyruvate (µM), and cerebral lactate-to-pyruvate ratio (LPR) in our software Odin [19]. The mean daily (24 h) and 3-day-period (72 h) values of ICP, CPP, and PRx55-15 values were calculated using minute-by-minute data. The corresponding mean arterial glucose values were based on ABG analyses from approximately every fourth hour. The mean values of cerebral glucose, cerebral pyruvate, and cerebral LPR were based on hourly analyses. All data were transferred to SPSS version 25 (IBM Corp, Armonk, NY, USA) for further statistical analysis. The Shapiro–Wilk test was used to test the normal distribution assumption. PRx55-15 showed normal distribution at all time intervals except day 2, whereas arterial glucose, cerebral glucose, cerebral pyruvate, cerebral lactate, cerebral LPR, and GOS-E did not show normal distribution. Hence, the Spearman correlation test was used to investigate the association between mean daily values of arterial glucose, PRx55-15, cerebral glucose, cerebral pyruvate, cerebral lactate, cerebral LPR, and outcome. There were only a few missing observations and those were excluded from the analyses. A p value < 0.05 was considered statistically significant. Many of the correlation tests were dependent, such as arterial glucose versus pyruvate, lactate and LPR, and correction tests which are particularly conservative in this case. As this was an exploratory study with preclinical support, we did not adjust for multiple comparisons, to avoid type II-errors. A multivariate linear regression analysis was done in the first subsequent 3 days post-injury to predict PRx55-15 based on age, Glasgow Coma Scale motor (GCS M) at admission, and mean ICP, CPP, and arterial glucose for each day.

All procedures performed in the studies involving humans were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained by the relatives of all participating patients.

There were 120 TBI patients included in this study. The mean age was 43 (± 20) years and 75% were male (Table 1). Median GCS M was 5 (interquartile range (IQR) 4–5). Outcome data showed a mortality rate of 10% and median GOS-E was 5 (IQR 3–7).

The mean ICP was 11 (± 8.5) mm Hg and the mean CPP was 75 (± 12) mm Hg on the first day post-injury (Table 2). The pressure autoregulatory status (PRx55-15) was worse on the first day of trauma with a mean value at 0.22 (± 0.21), but gradually improved to 0.16 (± 0.22) on the third day.

The mean arterial glucose was 8.8 (± 2.1) mM on the first day of trauma and decreased the consecutive 2 days, to 8.0 (± 1.3) mM on day three. The percentage of blood gases with hypoglycemic events, defined as arterial glucose below 5 mM, was at the mean value 1.7 (± 7.1) % on the first and 0.98 (± 5.6) % on the third day post-injury. The percentage of blood gases with hyperglycemic events, defined as arterial glucose above 10 mM, was at the mean value 19 (± 27) % on the first day post-injury and the mean value was 6.6 (± 15) % on day three.

The mean cerebral glucose decreased to some extent from 2.6 (± 1.3) to 2.1 (± 1.1) mM from days 1 to 3. The cerebral lactate was stable over the first 3 days, the mean value was 3.6 (± 2.5) mM on the first day of trauma. The cerebral pyruvate was 144 (± 90) µM on day 1 and 144 (± 52) µM on day 3 post-injury. The mean cerebral LPR gradually increased from 27 (± 20) on days 1 to 36 (± 86) on day 3. One patient had a LPR above 100 on day 1 (LPR = 181), three patients on day 2 (LPR = 134, 152, and 223, respectively), and two patients on day 3 (LPR = 463 and 758, respectively) (Fig. 1). The patients with high LPR on day 3 had developed intracranial hypertension above 50 mm Hg. There were no negative values; the standard deviation was high due to these outliers.

Sixty-eight of the 120 patients had at least one LPR value above 25 on the first 3 days. The cause of the pathologically high LPR was considered ischemic, when the concurrent cerebral pyruvate was below the ischemic threshold at 120 µM [20, 21], otherwise more indicative of a mitochondrial dysfunction. Cerebral pyruvate was below the ischemic threshold 120 µM in median 33 (IQR 0–84) % of the times LPR was above 25 in these 68 patients. Hence, cerebral pyruvate was more commonly normal to high at times of metabolic disturbances with LPR above 25, indicating mitochondrial dysfunction in these cases.

Arterial glucose and GCS M at admission did not correlate significantly on any of the first 3 days post-injury. However, mean arterial glucose correlated positively with mean PRx55-15 days 1 to 3 (r = 0.308, p value = 0.001), i.e., high arterial glucose with high PRx55-15/disturbed pressure reactivity (Fig. 2). Arterial glucose and PRx55-15 also had positive correlations with p values < 0.05 on days 2 and 3 (Table 3). Arterial glucose and cerebral LPR were associated days 1, 2 and 1–3 with p values < 0.05 (Fig. 3). Cerebral LPR and PRx55-15 had a positive association day 2 (r = 0.219, p = 0.048). Cerebral glucose (Fig. 4), cerebral pyruvate, and cerebral lactate had no correlation with PRx55-15 nor arterial glucose.

Four multivariate linear regression analyses were done to predict PRx55-15 on each of the first 3 days and the entire 3-day-period post-injury (Table 4). All analyses included age, GCS M at admission and the mean ICP, CPP, and arterial glucose of the day. Arterial glucose was a significant predictor of PRx55-15 on both days 1 and 2, but not on day 3. Increasing values of arterial glucose predicted higher PRx55-15/worse autoregulation these first 2 days. Increasing age predicted poor pressure autoregulation on day 3, and increasing ICP predicted worse pressure autoregulation on both days 2 and 3. Exclusion of those two patients that developed intracranial hypertension above 50 mm Hg attenuated ICP to be only a marginally significant predictor of PRx55-15, but did not otherwise have any major impact on the other coefficients.

Clinical outcome, defined as GOS-E, correlated significantly with the mean values of PRx55-15 days 1–3 (r = − 0.389, p value < 0.001), i.e., high PRx55-15 was associated with poor outcome (low GOS-E) (Table 5). There was also a negative association between arterial glucose days 1–3 and GOS-E (r = − 0.201, p value = 0.004), i.e., high arterial glucose with poor outcome (low GOS-E). The correlation between mean arterial glucose day 1 with GOS-E decreased from (r = − 0.381, p = 0.001) to (r = − 0.203, p = 0.045) on day 3 post-injury. The percentage of ABG with hyperglycemia (arterial glucose > 10 mM) only correlated with poor outcome on day 1 post-injury (r = − 0.251, r = 0.036). No associations could be detected between cerebral glucose, cerebral pyruvate, cerebral lactate, cerebral LPR, and clinical outcome.