The Neurochemical Signature of Cardiac Arrest: A Multianalyte Online Microdialysis Study

Figures

• 

  Figure 1. Experimental setup (A) Illustrative schematic of the experimental setup in the case of the one probe configuration. The anesthetized pig was placed on the operating table. The MD probe was inserted into the brain, and perfused with sterile perfusion fluid at 1 μL/min. The MD probe was inserted on average 3 h before baseline measurement commenced. (i) The outlet of the MD probe was connected to our wireless integrated neurochemical system, which consisted of a CMOS chip coated with a potassium ion selective membrane (ISM) and of a biosensor flow cell housing a glutamate, a glucose and a lactate biosensor. The dialysate concentration was monitored continuously and in real time throughout the experiment. The potentiometric CMOS chip (potassium) and amperometric biosensors (glucose, lactate and glutamate) communicated via Bluetooth with custom-made apps running on a portable laptop and tablets, respectively. (ii) Baseline levels were monitored for at least 30 min prior to commencing the experimental protocol. The experimental protocol began by inducing cardiac arrest followed by a 5 min “no-touch” period. The advanced life support protocol was then followed consisting of CPR using a…

• 

  Figure 2. Baseline box plots. Box plots showing baseline dialysate levels of glutamate (pink, n = 10), glucose (red, n = 18), lactate (green, n = 16) and potassium (purple, n = 5) at 1 μL/min, measured as the mean of the signal in the 5 min window immediately before the start of the cardiac arrest protocol. Box plots show the median and interquartile range and whiskers represent the 10th and 90th percentiles. For glutamate, the upper whisker is truncated to zoom in on the box plot as one value was much higher than the other values. Actual points overlaid on top of box plots.

• 

  Figure 3. Brain dialysate glucose and lactate traces during the cardiac arrest intervention protocol. (A) Example of the brain dialysate glucose level (red), lactate level (green) and the lactate/glucose ratio (blue) during the protocol. The data were sampled at 10 Hz, smoothed with a Savitzky–Golay 201-point filter and shifted to align time 0 with the time of cardiac arrest. The vertical line indicates the time of ROSC. At the end of the protocol the pig was euthanized with an infusion of potassium chloride (KCl). The horizontal dotted lines indicate the levels taken as baseline for glucose (red) and lactate (green) in this case. This was calculated as the mean of the signal in the 5 min window immediately before the start of the cardiac arrest protocol. (B) Box plots showing the percentage change in glucose (red, n = 16) and lactate (green, n = 14) from baseline after cardiac arrest. Glucose traces always showed a clearly defined trough following cardiac arrest, so levels were taken when they were at a minimum. For lactate, as the increase was less clearly defined between pigs, we chose to evaluate the level at 6 min after cardiac arrest to enable comparison. The 0% (horizontal dotted…

• 

  Figure 4. Brain dialysate glutamate response during the cardiac intervention protocol. Two examples of the brain dialysate glutamate response during the cardiac intervention protocol representing the two trends observed. The data were shifted to align time 0 with the time of cardiac arrest. Each vertical line indicates the ROSC time for the case associated with the same color. In one case (magenta) there was a transient peak in glutamate in response to cardiac arrest, whereas in the other (light pink) the glutamate level remained unchanged. The increase at 2.5 h after cardiac arrest in both traces corresponds to death. The two sets of data were sampled at 10 Hz, smoothed with a Savitzky–Golay filter (201-point for magenta and 1501-point for light pink).

• 

  Figure 5. Overlay of the normalized brain dialysate potassium traces following cardiac arrest. The percentage change in potassium from baseline is shown for five different cases. The data were shifted to align time 0 with the time of cardiac arrest. Each vertical line indicates the ROSC time for the case of the same color. ROSC was not achieved in 3 cases (light purple). The data were sampled at 0.3 Hz and were filtered using a Savitzky–Golay filter (25-point).

• 

  Figure 6. Online microdialysis measurements compared with discontinuous measurements. (A) Dialysate percentage change from baseline for glucose (red) and lactate (green) measured using the continuous online analysis system compared with offline analysis of discontinuous samples. Vertical lines indicate key time points in the protocol. (B) Another example of percentage change from baseline for glucose measured using the continuous online analysis system compared with discontinuous offline analysis for a second pig. In all cases the online MD probe was perfused at 1 μL/min and the offline probe was perfused at 2 μL/min. Any difference in probe recovery due to the different flow rates for the two methods is removed in the percentage change. Discontinuous offline measurements are represented by horizontal dotted lines, indicating the time that measurement was collected over.

• 

  Figure 7. Comparison of glucose levels in the dialysate with the net change in glucose between arterial blood samples and venous blood samples. The continuous MD data for glucose (in red) is compared with the changes in glucose in the blood samples that were taken at regular intervals during the clinical procedure and measured using a blood gas analyzer: at baseline, at the start of CPR, at ROSC, and afterward every 20 min until the end of the procedure. In more detail, the markers represent the difference between the glucose in the arterial blood sample and the venous blood sample, either mixed venous or cerebral venous, for each time frame. A positive value refers to a release of glucose from the tissue into the blood, while a negative value refers to an absorption of glucose into the extracellular space from the blood.