1.2.1.

Monitoring of NADH: history, principles, and basic technology

Since the discovery of pyridine nucleotides by Harden and Young,⁴⁵ more than 1000 papers have been published on the use of NADH as a marker for mitochondrial function. This type of research was started in 1955 by Chance & Williams^{46, 47} by defining the mitochondrial metabolic state in vitro. In order to understand mitochondrial function in vivo and under various pathophysiological conditions, it is important to monitor the redox state of the respiratory chain in real time

The discovery of the optical properties of NADH led to very intensive research since the early 1950s. The establishment of the method to monitor NADH in vivo was done between 1952 and 1962. An intensive use of the in vivo NADH monitoring approach began in 1962. The “classical” paper on in vivo monitoring of NADH was published in 1962 by Chance ³⁸ They were able to simultaneously monitor the brain and kidney of anesthetized rats using two microfluorometers. In 1962, Chance ⁴⁸ elaborated on this kind of in vivo monitoring and used it in other rat organs. The correction of the NADH fluorescence signal for changes in tissue absorption and scattering was initiated in the 1960s and 1970s by various investigators.^{49, 50, 51} Most of the published material in this field of in vivo NADH monitoring of tissues used the subtraction of the reflectance signal (366 nm) from that of the NADH fluorescence signal measured to provide the corrected fluorescence signal.^{52, 53} A few attempts were made in order to improve the correction technique, but no significant improvement was found. In addition, the various correction techniques were listed and discussed in detail in Ince ⁵⁴ Most of the published material was based on a 1:1 ratio, when subtracting the 366-nm reflectance from the 450-nm fluorescence signal. We have found that subtracting the reflectance from the fluorescence or dividing fluorescence over reflectance provides similar net NADH changes. In 2006, Bradley and Thorniley⁵⁵ reviewed the available techniques for correction of fluorescence signals. They concluded, “even though research has been conducted into correction techniques for over thirty years, the development of a successful and practical correction technique remains a considerable challenge.”

1.3.1.

Systemic pulse oximetry

This standard of care device is measuring the oxygenation level of the hemoglobin dependent on the pulse passing in the large arteries. In small vessels, the pulse disappears and, therefore, this technique cannot provide information on hemoglobin saturation in the microcirculation. Because of local changes in oxygen consumption, the systemic pulse may be relatively stable but the local Hb saturation may differ.

1.3.2.

Microcirculatory Hb oxygenation

The level of oxygenated hemoglobin can be monitored at the microcirculatory level using the absorption spectrum of hemoglobin, which is different in its oxygenated or deoxygenated state. One possibility is to illuminate the tissue with light at the isosbestic wavelength point (not sensitive to the oxygenation level of hemoglobin), and a second wavelength, at a nonisosbestic point, in which the oxyhemoglobin absorbs more light than the deoxyhemoglobin form. By subtracting the two-wavelength reflectance, a parameter correlated to blood oxygenation level is obtained.^{5, 56}

1.3.3.

Microcirculatory blood flow

Real-time monitoring of TBF can be achieved using laser Doppler flowmetry (LDF).^{57, 58, 59} The LDF measures relative changes, which correlate with the relative TBF alterations. The principle of LDF is to measure the Doppler shift, namely, the frequency change undergone by the light reflected from moving red blood cells. The results are presented as percentages of a full scale (0–100%), providing relative perfusion values.

1.3.4.

Mitochondrial NADH/flavoproteins fluorometry and near-infrared spectroscopy

NADH monitoring from the organ surface (brain, kidney, liver, testis, etc.) is performed by the fluorometric technique based on the original work by Chance and Williams.^{46, 47, 60} The excitation light (366 nm) is passed from the fluorometer to the tissue through a bundle of quartz optical fibers. The emitted light (450 nm), together with the reflected light at the excitation wavelength, is transferred to the fluorometer through another bundle of fibers. The measured changes in fluorescence and reflectance signals are calculated as percent values relative to the calibrated signals under normoxic conditions. This type of calibration is not absolute but provides reliable and reproducible results for different animals and different laboratories.^{41, 42}

Monitoring of flavoprotein (Fp) fluorescence in blood-perfused organs in vivo is very problematic as compared to blood-free organs in vitro. Two factors affect the very weak fluorescence signal. Blood oxygenation and blood volume changes in the microcirculation induced artifacts in the measurements, which are very hard or impossible to be corrected.

Near-infrared spectroscopy (NIRS) is a noninvasive optical technique that evaluates the relationship between oxygen availability and consumption by monitoring of the oxidation-reduction state of cytochrome aa3 (Cytaa₃), although the artifacts of the hemoglobin signal in the Cytaa₃ signal remain problematic. Because tissue oxygenation is indicated by the state of cytochrome oxidase, the terminal enzyme of the respiratory chain, dynamic measurements of oxygen utilization, and perfusion can be achieved by NIRS.^{61, 62}

1.3.5.

Microdialysis

The technique of tissue microdialysis allows continuous on-line monitoring of changes in tissue energy-related metabolites (e.g., glucose, lactate, pyruvate, adenosine, xanthine).⁶³ It involves the insertion of a catheter lined with a polyamide dialysis membrane into brain parenchyma that is perfused with a physiological solution (e.g., Ringer's lactate) at ultralow flow rates using a precision pump. Small molecules diffuse from the extracellular space into the perfusion fluid, which is then collected into vials that are changed every 10 to 60 min, allowing for up to 70% equilibration across the dialysis membrane.^{64, 65} Cerebral microdialysis has great potential for exploring the pathophysiology of acute brain injury, drugs pharmacokinetics within the central nervous system, and responses to therapeutic interventions.

1.3.6.

Tissue pO₂, pCO₂, and pH

Measurements of tissue pO₂, pCO₂, and pH are used as a monitoring modality in the neurosciences and critical care units,^{66, 67, 68} and as a marker of tissue oxygenation in research protocols.^{69, 70, 71} The technique involves the insertion of a microsensor into the brain parenchyma, either through a bolt inserted into the skull or directly through the craniotomy site and tunneled under the skin. At least two commercial microsensors allow direct, continuous measurement of brain tissue gases.⁷²

The aims of this paper are as follows:

2.1.1.

Microcirculatory blood flow

The optical fibers of the laser Doppler flowmeter were used for monitoring TBF. The LDF measures relative changes (in the 0–100% range), which are significantly correlated to the absolute values of TBF measured by other quantitative methods: H₂ clearance,^{58 14}C iodoantipyrine,⁵⁹ and radioactive microspheres.⁷⁴

2.1.2.

NADH redox state fluorometer/reflectometer

The principle of NADH monitoring from the tissue surface was described in Sec. 1.2.4. It is known that reflectance at 366 nm is inversely correlated to changes in the tissue blood volume;^{5, 39, 75, 76, 77, 78} therefore, the reflectance can be used as a hemodynamic parameter. The NADH fluorescence trace was corrected for hemodynamic artifacts by subtracting the reflectance signal (at 366 nm) from the fluorescence (450 nm) signal.^{42, 77} A detailed discussion about this issue can be found in our review articles.^{39, 41} The changes in the fluorescence and reflectance signals are calculated relative to the calibrated signals under normoxic conditions (100% NADH).^{41, 79, 80, 81, 82}

2.1.3.

Hemoglobin oxygenation reflectometer

Oxyhemoglobin measurement is based on the differences in the absorption properties of hemoglobin in its oxygenated versus deoxygenated states. The tissue is illuminated at two wavelengths: 585 and 577 nm. At 585 nm, oxyhemoglobin and deoxyhemoglobin have the same absorption characteristics (the isosbestic point); thus, the light emitted from the tissue reflects changes in tissue blood volume. At 577 nm, oxyhemoglobin has higher absorption ability (lower reflectance) than deoxyhemoglobin; thus, at this wavelength, the emitted light intensities are affected by the oxygenation levels as well as by the changes in blood volume at the measurement site. Therefore, subtracting changes in 585 nm reflectance from those at 577 nm provides a parameter correlated to net changes in blood oxygenation.^{5, 83}

2.2.1.

Light source unit

The LSU [see Fig. 6a] of the CRV comprises a 785-nm continuous wave (cw) laser diode, which serves for laser Doppler measurement; a UV light-emitting diode [(LED), 375 nm] for NADH fluorescence excitation and for total back scatter (or reflection) measurement; and a blue LED (470 nm) and green LED (530 nm) for HbO₂. In order to enable a very high measurement dynamic range of fluorescence and reflection parameters, the light source unit is designed to enable a very wide range of the excitation intensities. To enable this wide excitation range and linearity, the system is designed according to a three-floors concept. Each floor comprises three LEDs: one UV LED with emission peak at 375 nm, one blue LED with emission peak at 470 nm, and one green LED with emission peak at 530 nm. The different wavelengths from all LEDs of the same floor are assembled together and coupled into a single fiber by a set of dichroic mirrors and appropriate collimation and focusing lenses. The current of each one of the discrete LEDs is set by the appropriate electronics drivers directly controlled by a digital to analog (D/A) of the digital signal processor (DSP). The difference between the floors is the output intensity. There is a high-, medium-, and low-intensity floors. The different excitation intensities are achieved by utilizing various pinholes while maintaining all other electro-optical properties as the same for all three floors. The light from all three floors and a laser diode is combined into single mixer fiber therefore enabling precise setting of the excitation intensity within a very wide excitation range. The NIR laser diode at 785 nm, for laser Doppler measurements, operates in cw operation mode. The UV LEDs, blue LEDs, and green LEDs operate in chopping mode. This enables usage of synchronous detection techniques in order to detect the NADH fluorescence and total backscatter light. Additionally, the chopping operation mode enables one to perform NADH measurements with very low excitation intensities well below the limits specified by the laser safety standards.

2.2.2.

Detection unit

All six collection fibers of the fiber-optic probe are assembled into a single male SC optical connector. The light from the probe passes though the panel connector into a single thick optical fiber that delivers the light to the detection unit (DTU). At the DTU entrance, the collimation lens collimates the fiber output light. The collimated light is split according to the different wavelengths into the respective photodetectors by means of dichroic beam splitters. The first dichroic beam splitter reflects the total backscatter signal at 375 nm toward the photodiode detector. The higher wavelengths pass through the first beam splitter toward the second dichroic beam splitter. The second dichroic beam splitter reflects the NADH fluorescence signal at 450 nm and total backscatter signals at 470 and 530 nm toward the photomultiplier detector. Because of the chopping operation of the LEDs, the photomultiplier detector detects each one of the above-mentioned signals at different times (i.e., time-sharing operation detection mode). The second dichroic filter enables the laser Doppler signal at 785 nm to pass through it toward the photodiode detector. All acquired signals are digitized into the DSP by high-resolution 16-bit analog to digital (A/D).

2.2.3.

Digital signal processor

The DSP is responsible for whole system control, initial data processing, and calculation of the Doppler parameter. The DSP is built around a Tern Inc. 586-Engine-P controller board with AMD SC520 CPU. After initial data processing, the calculated values are transmitted to the panel computer for final data processing display through RS-232 serial interface. The CRV device utilizes medical-grade main power supply for all electronic circuits, including the panel computer.

2.2.4.

CritiView optical probes

Two types of fiber optic probes were developed for experimental animal and for clinical applications. (i) The pencil-style optical probe has one measurement point to be used on exposed tissue. The excitation and collection fibers are held together in a stainless steel element. It is comprised of one excitation fiber connected to respective excitation connector on the CRV panel and six collection fibers connected to the single collection connector on the CRV front panel. (ii) The Foley catheter optical probe for measuring the urethral wall vitality is based on a standard 3-lumen Foley catheter, which is an inflatable balloon retention-type catheter inserted through the urethra and used to drain the bladder. This probe construction enables the measurement of tissue metabolism at the urethral wall while draining urine from the bladder. The CRV Foley catheter optical probe is designed to perform optical measurements from three adjacent points on the tissue, 3.5-mm apart (Fig. 6). Each measurement point is comprised of one excitation fiber and two adjacent collection fibers. The three excitation fibers from the LSU are connected to the excitation connector on the CRV panel. All six collection fibers from the three measurement points are connected to the single collection connector on the CRV front panel. Figure 7 shows the finished probe inserted inside the three-way Foley catheter.

Fig. 6

(a) Schematic presentation of the CRV components. (b) Responses to ischemia and anoxia are presented. The effects of (left) ischemia and (right) anoxia on the four parameters measured from the brain of an anesthetized Gerbil. N₂, 100% nitrogen; R_occl, L_occl. occlusion of the right or left common carotid artery; R+L_open, reopening of the two carotid arteries.

Fig. 7

(a) The CritiView device used in the present study and (b,c) location of the three-way Foley catheter and the monitoring sites in the urethral wall.

2.4.1.

Animal preparation

In order to test the depth of anesthesia in each gerbil monitored, we used a surrogate systemic parameter that represents the function and the intactness of the cardiovascular and respiratory systems. The parameter is the level of the hemoglobin oxygenation in the systemic blood, SpO₂ measured by a standard veterinary pulse oximeter suitable for small animal experiments (model 8600V, Nonin Medical Inc., Minneapolis, Minnesota) The SpO₂ was very high and stable during the entire period of monitoring. During the four perturbations imposed on the gerbil, a clear decrease of the SpO₂ was recorded. This decrease in SpO₂ during anoxia, hypoxia, and terminal anoxia is according to the availability of oxygen to the body. The decrease in SpO₂ during cerebral ischemia is probably due to stimulation of the sympathetic and parasympathetic nerves that are affected by the pulling of the carotid arteries during the ischemic episode.

2.4.2.

Operation procedure

Male rats (250–300 gr) and mongolian gerbils (50–75 gr) were used. We had operated 35 rats and 25 gerbils that were used for the various in vivo studies. This includes the testing of the various devices as well as the comparison studies between the TVA and the CRV. The animals were anesthetized by Equithesin (E-th = chloral hydrate 42.51 mg; magnesium sulfate 21.25 mg; alcohol 11.5%; propylene glycol 44.34%; pentobarbital 9.72 mg) intraperitoneal (IP) injection of 0.3 ml/100 g body weight. The animals were kept anesthetized during the operation, and during the entire monitoring period, by additional IP injections of E-th 0.03 ml in gerbils and 0.1 ml in rats, every 30 min. The addition of small volumes of E-th every 30 min kept the animals in a stable state. In the gerbils, the two common carotid arteries were isolated just before brain surgery and ligatures of 4–0 silk thread were placed around them. The animal was placed in a head holder in the supine position. After a midline incision of the skin, an appropriate elliptic hole (about 3×8 mm) was drilled in the parietal bone of the right hemisphere. The dura mater remained intact, and a light guide holder—cannula—was placed in the drilled hole, and extra pressure on the tissue was avoided. Two stainless steel screws in the left parietal bone were used to fix the cannula, with dental acrylic cement. Body temperature was measured by a rectal probe (Yellow Springs Instrument, Yellow Springs, Ohio) and was regulated to be at the range of 35–37°C using a heating blanket.

2.4.3.

Metabolic perturbations

Anoxia. Exposure of the animal to oxygen deficient atmosphere by spontaneous breathing of 100% N₂, for a period of time (∼25 s) until changes are observed on the monitor and the animal has stopped breathing. The animal is then allowed to breathe normal room air by spontaneous breathing or by short artificial respiration.

Ischemia: Reversible occlusion (30 s) of the two common carotid arteries by constricting them with threads.

Hypoxia: Exposure of the animal to spontaneous breathing of low oxygen concentration (6% oxygen in air) for a short period of time (60 s). The animal is then allowed to breathe normal room air by spontaneous breathing.

Terminal Anoxia: Exposure of the animal to complete irreversible anoxia by exposing the animal to 100% N₂ until the death of the gerbil.

The application of the various gases was done by placing a small tube, connected to the gas cylinder, around the nose of the animals. The flow of the gas was very slow (<1 l/min) in order to avoid respiratory disturbances.

2.4.4.

Experimental protocol

Monitoring the various parameters was started immediately after the end of operation. This was done by using two needle-type optical probes connected to the CRV and the CritiView-CRV1 and placed inside the cannula cemented to the skull. In order to check the intactness and the physiological status of the brain, a 20-s anoxia was induced. Rats and gerbils that present abnormal response were discarded before running the experimental protocol. In each monitoring system, the variation in the response to short anoxia was in the range of ∼10%. For example, if the increase in NADH was in the range of 50–60%, any animal that showed 20–30% or 70–80% was not involved in the study and discarded at this stage of the study. Normally, only 5–10% of the operated animals were discarded at this stage. When body temperature reached the range of 35–37°C, the exposure of the animals to the various perturbations started. The animals were exposed to the various conditions having an interval of ∼10 min between each test to allow recovery. The pulse oximeter value was used in order to assure the complete recovery. The data were collected at a rate of one sample per second and stored in different channels of a computerized data acquisition program.

3.3.1.

Patients in the intensive care unit

Three responses measured in ICU patients are shown in Fig. 10. We assumed that, under stressful situations, in a less vital organ (such as the urethral wall) the microcirculatory blood flow will be reduced and oxygen delivery to the mitochondria will diminish. In a female patient [Fig. 10a], the responses to a cessation of spontaneous respiration were recorded. As seen, a clear hypoperfusion was recorded in the urethral wall together with a significant increase in NADH levels. As soon as artificial ventilation was restarted, the signals returned to baseline levels. In another patient, a dramatic decrease in the urethral wall energy state was recorded during a standard airway-suction procedure. As Fig. 10b shows, a large decrease in the microcirculatory blood flow was measured simultaneously with a marked increase in NADH. These changes suggest that the cardiorespiratory stressful situation induced by suction could be detected by the decrease in the urethral wall viability.

Fig. 10

The results of three patients monitored in the ICU using an early version of the device named the tissue spectroscope: (a) Responses to a cessation of spontaneous respiration in a female patient, (b) effects of the standard airway suction procedure on the viability of the urethral wall, and (c) monitoring of a patient in the cardiac ICU during the initial 4 h after the bypass surgery, showing the development of cardiac tamponade.

The third patient was monitored during the early postoperative period after coronary artery bypass graft (CABG) operation [Fig. 10c]. The left side of Fig. 10c shows that, during the initial 3 h of monitoring, the NADH levels were relatively stable (±20% change). During the next 40 min of monitoring [right side of Fig. 10c], a marked mitochondrial dysfunction was recorded (100% increase in NADH) in parallel to a decrease in TBF. The NADH response included two types of change, namely, a pronounced continuous increase in the NADH levels as well as a number of transient elevations lasting 1–5 min each. It was found that, during this period, a cardiac tamponade was developed and the patient was returned to the operating room for resternotomy. The hemodynamic parameters (heart rate and blood pressure), shown in the boxes below Fig. 10, were not clearly correlated to the development of the tamponade.

3.3.2.

Patients undergoing open repair of abdominal aortic aneurysm

In five patients, clear but not entirely consistent responses to the clamping of the abdominal aorta were observed and recorded. During this phase of the operation, the microcirculatory blood flow in the urethra decreased in all patients, while the results of NADH measurements showed some variability. Typical responses to aortic occlusion in one of the abdominal aortic aneurysm (AAA) patients are shown in Fig. 11a. The preparation of the aorta for the clamping procedure (marked as A) led to a clear transient decrease in TBF, as well as an increase in the NADH level. The clamping of the aorta led to a maximal decrease in TBF in parallel to the increase in NADH. In this patient, the monitored HbO₂ showed a clear decrease during the clamping interval. Because of a decrease in tissue blood volume, during the aortic occlusion, the reflectance trace showed a large increase until the reopening of the occluded aorta. Immediately after declamping, the initial small increase in TBF led to a fast recovery of the HbO₂ and NADH redox state. All signals recovered to the same values measured during the short control period; although, the occlusion interval lasted for >80 min. In Fig. 11b, the typical correlation between the various parameters is presented. Data were extracted from the five patients of this group according to the six stages of the operation: Stage 1, 10 min before the clamp; stage 2, 10 min after the clamp; stage 3, midpoint of the clamping period; stage 4, 10 min before the declamp; stage 5,- 10 min after the declamp; and stage 6, the last 5 min of monitoring in the OR.

Fig. 11

(a) Effects of aortic occlusion in a patient that underwent an AAA repair operation, on the four parameters monitored by the CritiView. The technical preparation for aortic occlusion was performed at the marker A. (b) the correlations between TBF and NADH/reflectance or HbO₂ and NADH. (c) The Mean ± SE changes of TBF and NADH in five patients who underwent the AAA operation.

The results for the five patients [Fig. 11c] showed a pronounced decline in the mean TBF values, noted in our experiments from the second stage of the surgery up to the fourth stage. The dependence of TBF on time was statistically significant. The difference was significant for all stages of surgery (t1–t4, p = 0.004), but was non-significant (p > 0.05) for the partial intervals (t2–t3 and t3–t4). This analysis indicates that the only statistically significant change is between points 1 and 4. The changes in the NADH level were significant only when the basal level (t1) was compared to t3.

3.3.3.

Patients undergoing open heart surgery

Thirteen patients underwent a coronary artery bypass using a cardiopulmonary bypass. Two other patients underwent an off-pump coronary artery bypass. Another two patients had valve procedures, and two underwent aortic repair using a cardiopulmonary bypass. The last patient was exposed to aortic repair as well as the CABG procedure. Therefore, 18 out of 20 patients were operated with the use of the cardiopulmonary bypass heart-lung machine and two were operated under beating heart conditions. Two out of the 20 patients were exposed to deep hypothermia. Three typical responses to the cardiovascular operation procedure are presented in Fig. 12.

Fig. 12

Effects of the surgical procedure in (a–c) three cardiovascular operated patients on the urethral TBF and NADH redox state (see text for detailed explanations) and (d) the Mean ± SE changes of TBF and NADH in 11 patients during heart bypass operation.

In the first patient [shown in Fig. 12a], the responses of TBF and NADH were recorded as soon as the preparation for operation started with the scrubbing of the chest area (marked as A). This very early response was found only in one patient monitored in the cardiovascular operating room. The low level of TBF and the high level of NADH were recorded during the entire operation procedure, and the recovery began as soon as the chest was closed. When the patient left the operating room for the cardiac ICU, the two parameters were close to the baseline levels. These fast responses are probably due to high sensitivity and a fast response of the autonomic nervous system. We could speculate that the level of adrenaline was elevated very early, and as a result, the microcirculatory blood flow was decreased dramatically and the mitochondrial NADH became highly reduced. In the second patient who underwent the routine CABG operation [Fig. 12b], the same type of response was recorded. TBF reduction and NADH elevation were observed during most of the surgical procedure. Spontaneous transient recovery of TBF and NADH were noted during the second half of the operative procedure. In the third patient, operated on for aortic repair [Fig. 12c], clear responses to the procedure were recorded. In this patient, at 16:49, the initiation of extracorporeal circulation led to a large decrease in TBF and a large increase in NADH. The signals reverted toward the initial values, although the baseline was not reached (the monitoring period ended at 18:14). In order to perform statistical analysis, we selected patients that underwent a similar operation procedure. Only 13 patients were exposed to the same surgical protocol, and data were extracted from 11 patients. In two patients from the CABG normothermic group, the data collection was interrupted in the middle of the operation and we were unable to calculate the results to match the other 11 patients analyzed. The other seven patients were exposed to a different protocol and were not included in the statistical analysis. The Mean ± S.E. changes of TBF and NADH were obtained in 11 patients [Fig. 13b] following heart bypass operation during seven stages of the operation: Stage 1, 5 min baseline (after insertion of the catheter); stage 2, 10 min after chest opening; stage 3, 10 min after HLM or CPB was on; Stage 4, midpoint of HLM; stage 5, 10 min before HLM was off; stage 6, 10 min after HLM was off; and stage 7, the last 5 min of monitoring. ANOVA multiple comparison tests were conducted on the first three stages of the operation. Significance was found between points 1 and 3 in the TBF and between point 1 and points 2 and 3 in NADH (*p < 0.05, **p < 0.01, n = 11).

Fig. 13

Responses of the urethral wall TBF and NADH to systemic hypothermia (17°C) in a patient who underwent the aortic repair operation: (a) Responses to bleeding under normothermia and (b) responses to a decrease of microcirculatory blood flow to the lower part of the body.

In one patient, it was possible to compare the responses of the urethral wall to ischemia induced under normothermic and hypothermic conditions. The results presented in Fig. 13 were measured in a patient who underwent aortic repair surgery. In the process of opening the chest, a major bleeding occurred and the patient was immediately put on the heart-lung machine and severe hypothermia was induced. As can be seen in Fig. 13a, while the patient was normothermic, large changes in TBF and NADH were noted during the major bleeding event. After cooling the patient and shifting him to the heart-lung machine, the signals recovered, although the TBF remained under the control values. While the patient was hypothermic (17°C), the microcirculatory blood flow was directed mainly to the brain area [Fig. 13b], whereas the TBF to the urethra was decreasing significantly without any clear corresponding change in the NADH levels. This may suggest that the mitochondrial redox state behaves differently under hypothermia. It should be acknowledged that this response was recorded only in one patient.