Short-term, Mild Hypothermia Can Increase the Beneficial Effect of Permissive Hypotension on Uncontrolled Hemorrhagic Shock in Rats

• Permissive hypotensive resuscitation is advocated for traumatic, hemorrhagic shock

• The combination of permissive hypotension and mild hypothermia in anesthetized rats increased their survival time and survival rate compared with normotensive and normothermic rats, all exposed to similar states of hemorrhagic shock

IT is estimated that 15–20% of traumatic deaths are potentially preventable, and that 66–80% of these deaths are from hemorrhagic shock.1,2Effective fluid resuscitation is the key step in follow-up therapy for hemorrhagic shock.3,–,5 

Based on optimal fluid resuscitation and transfusion to patients with hemorrhagic shock and life-threatening injuries, the concept of damage-control resuscitation has been developed.6,–,8The main principles are permissive hypotensive resuscitation, rapid definitive/surgical control of bleeding, prevention/treatment of hypothermia, acidosis, and hypocalcemia. The goal of damage-control resuscitation is to minimize iatrogenic resuscitation injury, prevent worsening of the presenting traumatic shock, and to obtain definitive hemostasis. Research has shown that permissive hypotensive resuscitation results in a good resuscitative effect for uncontrolled hemorrhagic shock.9Permissive hypotensive resuscitation is aimed to reduce blood loss, prevent the translocation of a forming thrombus, and rebleeding.10Our previous study11showed that 50–60 mmHg of mean arterial pressure (MAP) is an ideal resuscitative pressure during uncontrolled hemorrhage.

Damage-control resuscitation advocates prevention of a decrease in body temperature during resuscitation after shock,12but reports have shown that short-term and mild-to-moderate controlled hypothermic resuscitation is beneficial in the treatment of uncontrolled hemorrhagic shock.13,–,16Our previous study also demonstrated that short-term application of mild or moderate hypothermia to uncontrolled hemorrhagic shock improved the final resuscitative effect, whereas too low and long-term hypothermia did not improve the resuscitative effect. No study has investigated whether short-term and mild hypothermia results in beneficial effects of hypotensive resuscitation or elucidated the underlying mechanism. Based on other and our previous studies, we hypothesized that short-term and mild hypothermia during hypotensive resuscitation improves survival compared with hypotensive resuscitation, and protects organ function via  decreasing oxygen consumption and tissue metabolism.

To test this hypothesis, the effects of permissive hypotension (50 mmHg) in combination with mild hypothermia on uncontrolled hemorrhagic shock (as well as the effects on hemodynamics, coagulation parameters, vital organ and mitochondrial function, and energy metabolism) were observed.

The study protocol was approved by the Research Council and Animal Care and Use Committee of the Research Institute of Surgery, Daping Hospital, Third Military Medical University (Chongqing, China). None of the authors are members of this committee.

Two-hundred and twenty Sprague-Dawley rats (220–260 g) were used in the present study. They were fasted for 12 h, but allowed water ad libitum  before experiments. The number of animals used had to be sufficient so that we could discover if short-term and mild hypothermia during hypotensive resuscitation improved survival compared with hypotensive resuscitation. The sample number (n = 8 or 10) in each group was calculated by power analyses based on our previous study and results from pilot studies (α at 0.05, power at 80%, two-tailed).

On the day of the experiment, rats were anesthetized with sodium pentobarbital (30 mg/kg). This agent was then added until the rats had no response to a needle stimulus. Rats were spontaneously breathing without mechanical ventilation. The right femoral arteries and veins were catheterized with a polyethylene catheter (OD, 0.965 mm; ID, 0.58 mm) for monitoring the MAP and bleeding, respectively. Rats underwent left ventricular catheterization with an appropriately sized catheter via  the right carotid artery. To monitor hemodynamics and prevent clot formation, rats were given 500 U/kg of heparin and all catheters were filled with physiologic saline (0.9%) containing 30 U/ml of heparin. An uncontrolled hemorrhagic shock model was induced by transection of the splenic parenchyma and one of the branches of the splenic artery, as previously described by our research team.11Briefly, after the completion of catheterization, the spleen was exposed after laparotomy and a cross-transection was made in the splenic parenchyma between the two major branches of the splenic artery. One of the major branches of the splenic artery was also transected. Blood was allowed to flow into the abdominal cavity. When the MAP decreased to 40 mmHg, uncontrolled hemorrhagic shock was established for subsequent experiments.

Experiments were defined as four phases. Phase I was the uncontrolled hemorrhagic period (model stage) in which blood was freely hemorrhaged to the abdomen. This phase was achieved when the MAP decreased to 40 mmHg, and this period was maintained for approximately 20–30 min. Phase II was the hypotensive and hypothermic resuscitation period in which rats were maintained at different target MAPs (including 50 mmHg and 80 mmHg) in combination with mild hypothermia (34°C) or normal temperature (37°C) with infusion of 130 kD of hydroxyethyl starch and lactated Ringer's solution at a ratio of 1:2. Phase III was the definitive treatment period; after bleeding was completely controlled via  ligation of the spleen artery, rats received whole blood and lactated Ringer's solution (1:2) to maintain the MAP at 90–100 mmHg for 2 h. Phase IV was the observation period, in which the follow-through effects of hypotensive and hypothermic resuscitation were observed for 2 h. In phase II, rats in the mild hypothermia (34°C) resuscitation group underwent external application of ice on the head and abdomen to keep the rectal temperature at 34°C. (fig. 1).

Fifty Sprague-Dawley rats were randomly divided into five groups of 10 as follows: sham-operated; 50 mmHg at 34°C; 50 mmHg at 37°C; 80 mmHg at 34°C; and 80 mmHg at 37°C. Animal management and resuscitation regimens were described in the section of experimental phases. The animal number in each group was based on our previous work and power analyses. MAP, hemodynamic parameters (including left intraventricular systolic pressure, maximal change rate of left intraventricular pressure [±dp/dt^max]), and blood gases (including pH and partial pressure of oxygen and carbon dioxide in arterial blood [PaO², PCO²]) were determined at baseline, at the end of phase I (model phase), phase II (hypotensive and hypothermic resuscitation), phase III (definitive treatment stage), and 2 h during phase IV (observation period) with a polygraph physiologic recorder (SP844, Power Laboratory; AD Instruments, Castle Hill, NSW, Australia) and a blood gas analyzer (Phox plus L; Nova Biomedical, Waltham, MA). All hemodynamic parameters were obtained by Power Laboratory software, Chart V5. Blood loss, the amount of fluid infusion, and animal survival were recorded. The amount of blood loss was measured at the end of phase II using the method of cotton weighing.

In this experiment, another 50 Sprague-Dawley rats underwent catheterization of the left ventricle and right external jugular vein for measurement of cardiac output. The other procedures (including grouping, production of hemorrhage shock, and fluid infusion) were the same as described in experiment series 1. Each fluid tested at every target MAP involved experiments on 10 rats. Two hours after fluid infusion during phase III, cardiac output was measured using a Cardiomax-III machine (Columbus Instruments, Columbus, OH). Blood was sampled from the femoral artery and vein to analyze arterial and venous blood gases, respectively, using a blood gas analyzer. The values of VO²in tissues were calculated using the following formulas:

with VO²= oxygen consumption, CI = cardiac index, SaO²= oxygen saturation of the artery, and SvO²= oxygen saturation of the vein.

Eighty Sprague-Dawley rats were randomly divided into five groups (n = 16/group; each time-point, 8 rats) as follows: sham-operated; 50 mmHg + 34°C; 50 mmHg + 37°C; 80 mmHg + 34°C; and 80 mmHg + 37°C. Blood was sampled to measure the function of the liver and kidneys, and the level of D-lactic acid in the plasma at the end of phase II and III. Thereafter, rats were killed to remove the liver, kidneys, and small intestine for the measurement of mitochondrial function and energy metabolism. The variables of liver function (total bilirubin), kidney function (blood urea nitrogen), and serum creatinine), as well as the variables of hepatic cell damage (alanine aminotransferase and aspartate aminotransferase), were measured using a biochemical analyzer (DX800; Biochemical Analyzer, Beckman, Fullerton, CA). The intestinal barrier function was reflected by plasma D-lactic acid, which was determined by an enzymatic technique.17 

Mitochondrial function was measured by a mitochondrial function analyzer (MT 200; Strathkelvin, Lanarkshire, Scotland). Mitochondrial function reflected the respiration control rate of mitochondria, which is the rate of consumed oxygen of mitochondria with and without adenosine diphosphate. The respiration control rate is a relevant index to reflect the construction integrity and oxidative phosphorylation function of mitochondria. Energy metabolism was reflected by Na⁺-K⁺-ATPase, which was determined using an enzyme-linked immunosorbent assay.

Samples of liver or kidney tissue (5 g) were put into 20 ml of ice-cold isolation buffer (sucrose 0.25 M, Na²EDTA 0.1 mM, Tris 0.01 M, pH 7.6). They were cut into small pieces, and washed three times to remove blood. Tissue with isolation buffer was homogenized and then centrifuged at 1600 ×g  for 12 min at 4°C. The supernatant was further centrifuged twice at 25,000 ×g  for 15 min at 4°C. The pellet was collected and resuspended in 1 ml isolation buffer. The concentration of mitochondrial protein was measured by the Lowry method; 1.4 ml of measurement buffer (Tris 0.2 M, pH 7.6, KCl 15 mM, KH²PO⁴15 mM, Na²EDTA 1 mM, MgCl²5 mM, sucrose 0.25 M) warmed to 30°C was added into the reaction chamber and equilibrated for 2 min. Then 0.2 ml of 3 mg/ml of a mitochondrial mixture was put into the reaction chamber and equilibrated for 20 s. Ten microliters of 0.5 M sodium malate (C⁴H⁴Na²O⁵· H²O) and sodium glutamate (C⁵H⁸NNaO⁴) and 5 μl of adenosine diphosphate (400 nM) were added in succession. The rate of oxygen consumption was determined using the mitochondrial function analyzer.

Five grams of tissue from the liver, kidneys, or intestine were homogenized and centrifuged at 1,000 ×g  for 30 min. The supernatant was collected and the protein concentration measured by the Lowry method. Ten microliters of diluted testing sample was added into each well of a 96-well plate and incubated at 37°C for 30 min. Fifty microliters of horseradish peroxidase-conjugated reagent was then added to each well and mixed, and incubated for another 30 min at 37°C. Chromogen solution was then added to each well, and incubated for 15 min at 37°C. The optical density was measured at 450 nm after adding stop buffer with a Multiscan Spectrum 1500 machine (Thermo Electron Corporation, Waltham, MA). The activity of Na⁺-K⁺-ATPase in each sample (μmol/mg) was calculated by the linear regression equation of the standard curve based on the optical density value of the sample.

Forty Sprague-Dawley rats were randomly divided into four groups (n = 10/group): 50 mmHg + 34°C; 50 mmHg + 37°C; 80 mmHg + 34°C; and 80 mmHg + 37°C group. To avoid the interference of heparin on coagulation function, heparin was not used in this experiment. Blood was sampled to measure coagulation parameters such as the thrombin time (TT), prothrombin time (PT), international normalized ratio of prothrombin time (PT-INR), activated partial prothrombin time (APTT), fibrinogen, platelet count, and aggregation rate at baseline, at the end of phase I, II, III, and 2 h of phase IV.

Data (hemodynamic parameters, blood gas parameters, coagulation parameters, fluid and blood requirements, blood loss, hemoglobin, cardiac output, oxygen consumption [VO²], mitochondrial respiratory control rate, organ function parameters and Na⁺-K⁺-ATPase) are presented as the mean ± SD of n observations. Fluid and blood requirements were analyzed by one-way ANOVA, followed by the post hoc  Tukey test (SPSS 15.0; SPSS, Inc., Chicago, IL). Hemodynamic parameters, blood gas parameters, coagulation parameters, blood loss, hemoglobin, cardiac output, VO², mitochondrial respiratory control rate, organ function parameters, and Na⁺-K⁺-ATPase were analyzed by two-way ANOVA, followed by the repeated post hoc  Tukey test. Animal survival time and survival rate were analyzed by median and interquartile range, Kaplan–Meier survival analyses, and the log-rank test. P < 0.05 was considered significant (two-tailed).

The mean fluid requirement (hydroxylethyl starch + lactated Ringer's, ratio of 1:2) in the hypotensive resuscitation group was less than in normotensive resuscitation group at normothermia (4.3 ml vs.  15.6 ml) during phase II (P < 0.01). Application of hypothermia in phase II further reduced the average fluid requirement of the hypotensive resuscitation group (reduced to 2.7 ml from 4.3 ml; fig. 2A). Blood requirements in the two hypotensive resuscitation groups during phase III was less than in the two normotensive resuscitation groups, but there were no differences among all four groups (P > 0.05) (fig. 2A).

The mean blood loss of uncontrolled hemorrhagic shock for each rat in phase I was 5.66 ± 0.44 ml (mean ± SD). Permissive hypotensive resuscitation (50 mmHg) during phase II slightly increased blood loss; blood loss was increased by only 0.53 ml and 1.94 ml in the 50 mmHg target MAP group at 34°C or at 37°C, respectively. Normotensive resuscitation (80 mmHg) increased blood loss as compared with hypotensive resuscitation (P < 0.01); blood loss was increased by 4.82 ml and 6.98 ml in the 34°C or 37°C group in normotensive resuscitation, respectively (fig. 2B).

As compared with the sham-operated group, the mean hemoglobin concentration was decreased at the end of phase I (from approximately 11.7 g/100 ml to approximately 8.8 g/100 ml). Hypotensive resuscitation only slightly decreased the concentration of hemoglobin, whereas normotensive resuscitation significantly decreased the concentration of hemoglobin during phase II (from approximately 8.8 g/100 ml to 2.5–3.4 g/100 ml) (P < 0.01). In the same resuscitation pressure groups, the concentration of hemoglobin in the hypothermia group was higher than in the normothermia group (fig. 2C).

As compared with the normotensive/normothermia resuscitation group, short-term and mild hypothermic/hypotensive resuscitation increased the survival time and survival rate of Sprague-Dawley rats. The number of rats surviving was 9/10, 8/10, 4/10, and 1/10 in the 50 mmHg at 34°C or 37°C and 80 mmHg at 34°C or 37°C groups at the 2 h time-point of phase IV, respectively. Two hypotensive resuscitation groups were significantly superior to the two normotensive resuscitation groups (P < 0.05 or 0.01). Mild hypothermia in combination with hypotensive resuscitation had the best survival time and survival rate (fig. 3).

MAP, left intraventricular systolic pressure, and ±dp/dt^maxin all groups at baseline and at the end of phase I were not different. During phase II, because of the difference in target pressure, these hemodynamic parameters in normotensive resuscitation groups were higher than in hypotensive resuscitation groups. However, during phases III and IV, these hemodynamic parameters in normotensive groups were lower than in the hypotensive resuscitation groups. Among all groups, hypotensive resuscitation in combination with mild hypothermia (50 mmHg at 34°C group) had the best hemodynamic parameters during phases III and IV (table 1).

There were no differences in blood pH, PaO², and PaCO²at baseline and at the end of phase I among all groups. At the end of phase II, the blood pH value in all groups decreased slightly as compared with baseline and at the end of phase I. During phase III and phase IV, the pH value in hypotensive resuscitation groups (irrespective of 34°C or 37°C) was restored to a normal level, which was higher than the two normotensive resuscitation groups (P < 0.05). PaO²values in the two hypotensive resuscitation groups were higher than the two normotensive groups at all time-points (phases II, III, and IV) (P < 0.05 or 0.01). PaCO²values in the hypotensive resuscitation groups were lower than in the normotensive groups at the end of phase II (P < 0.05). In the same resuscitation pressure groups, PaO²in the mild hypothermia group was higher than in the normothermia group (P < 0.05) (table 2).

In phase II, cardiac output in normotensive resuscitation groups (at 34°C and 37°C) was slightly higher than in the hypotensive resuscitation groups (at 34°C and 37°C) in the same target pressure group. That is, hypothermia resuscitation (34°C) had no impact on cardiac output. During phase III, cardiac output among all groups was not obviously different (P > 0.05) (fig. 4A).

In the normotensive resuscitation groups, VO²was higher than in the hypotensive resuscitation groups at the end of phase II (P < 0.01). Hypothermia decreased VO²(fig. 4B). There were no significant differences among all groups for VO²at the end of phase III (P > 0.05). (fig. 4B).

Hypotensive resuscitation in combination with mild hypothermia could better protect vital organ function (including intestinal barrier function). The plasma level of D-lactic acid, the concentration of total bilirubin, blood urea nitrogen, serum creatinine, alanine aminotransferase, and aspartate aminotransferase were significantly lower than in the normotensive/normothermia, normotensive/hypothermia, and hypotensive/normothermia groups, and almost close to the normal level (sham-operated group; figs. 5A–F). In the same target MAP group, the variables of the intestinal barrier, liver, and kidney function in the mild hypothermia group were superior to that in the normothermia group, whereas in the same temperature group, the hypotensive group was superior to the normotensive group.

Hypotensive resuscitation in combination with mild hypothermic resuscitation could better protect the mitochondrial function of the liver, kidneys, and intestinal mucosa. The respiratory control rates of mitochondria from these organs were close to the normal level (sham-operated group). Two hypotensive resuscitation groups (50 mmHg at 34°C and 37°C) were superior to the two normotensive groups (80 mmHg at 34°C and 37°C) (P < 0.05 or 0.01). In the same target pressure resuscitation groups, the mild hypothermia group (34°C) was superior to the normothermia group (37°C; figs. 6A–C) (P < 0.05).

The activity of Na⁺-K⁺ATPase in the liver, kidneys, and intestinal mucosa was increased after hemorrhagic shock as compared with the control group (sham-operated). Hypotension in combination with mild hypothermic resuscitation reduced the activity of Na⁺-K⁺ATPase in the liver, kidneys, and intestinal mucosa. In the same target MAP group, the activity of Na⁺-K⁺ATPase in the hypothermia resuscitation group was lower than in the normothermia group (figs. 7A–C) (P < 0.05 or 0.01).

At the end of phase I, TT, PT, and PT-INR did not change obviously as compared with baseline. Fibrinogen decreased slightly (P > 0.05). APTT was prolonged, platelet count was decreased, and platelet aggregation increased (P < 0.05). During phase II, TT, PT, PT-INR, and APTT were prolonged or increased as compared with the end of phase I (P < 0.05 or 0.01); fibrinogen, platelet count, and aggregation were decreased (P < 0.05 or 0.01). Among these parameters, PT, PT-INR, and APTT in the hypothermia group (34°C group) were slightly higher than in the normothermia group (37°C group) at the same pressure level (P < 0.05), and platelet aggregation in the hypothermia group was lower than in normothermia group. During phases III and IV, some parameters appeared to be slightly increased and some parameters appeared to be slightly decreased, but there were no significant differences in most parameters between hypothermia and normothermia groups (table 3) (P > 0.05).

Traumatic hemorrhagic shock occurs often in civilian and military settings. Traumatic hemorrhagic shock accounts for approximately 50% of deaths of battle personnel. Fluid resuscitation is the common and very important treatment for many types of circulatory shock (particularly for traumatic hemorrhagic shock).18A great deal of research has shown that permissive hypotensive resuscitation and short-term hypothermic resuscitation results in a good resuscitative effect for uncontrolled hemorrhagic shock before bleeding is controlled. Long-term and deep hypothermia do not achieve a good effect.19Whether short-term and mild hypothermia can produce a beneficial effect of hypotensive resuscitation before bleeding is controlled (and how it is enforced) are not understood.

The present study showed that permissive hypotensive resuscitation in combination with mild hypothermia had a good effect on uncontrolled hemorrhagic shock; short-term and mild hypothermia can potentiate this beneficial effect of hypotensive resuscitation. Permissive hypotensive resuscitation in combination with mild hypothermia can decrease the blood loss and functional damage of vital organs, decrease tissue oxygen consumption and metabolism, protect mitochondrial function, stabilize hemodynamic parameters, and maintain the acid–base balance of shocked animals (P < 0.05 or 0.01). The beneficial effect of permissive hypotension is mainly related to preventing blood loss and hemodilution. The beneficial effect of hypothermia is mainly related to decreasing metabolism and protecting organs.

The literature shows that subhypothermia has been successfully applied to cranial trauma patients.20However, application of hypothermia in trauma patients or hemorrhagic shock victims is controversial.21Some authors suggest that trauma patients are already hypothermic, and further application of hypothermia can interfere with cardiac function, the activity of many enzymes, and with coagulation.21,–,23However, other authors suggest that short-term and mild hypothermia would not significantly interfere with cardiovascular function and the activity of so many enzymes. In contrast, it may protect organ function by decreasing the metabolic rate.24,–,26The present study demonstrated that short-term application of mild hypothermia limited to before bleeding has stopped can decrease the energy metabolism, decrease the activity of Na⁺-K⁺-ATPase, decrease oxygen consumption, and protect mitochondrial and vital-organ function. However, it affected coagulation function only slightly. The present study further confirmed the efficiency of short-term and mild hypothermia in the early treatment of traumatic hemorrhagic shock. It also showed that hypothermia can increase the beneficial effect of hypotensive resuscitation on uncontrolled hemorrhagic shock.

There are many approaches to decreasing the body temperature of traumatic victims (e.g. , external application of ice or alcohol, use of a hypothermic bed, or intravenous infusion of cold fluid). For example, Takasu and Wu used alcohol and ice to implement hypothermic treatment on uncontrolled hemorrhagic shock in rats.27,28Wladis et al.  29used a hypothermia bed to observe the effect of hypothermia on acute metabolism and endocrine responses after hemorrhagic shock. Gotberg et al.  utilized intravenous infusions of cold saline to induce hypothermia to observe the beneficial effect on cardiogenic shock in pigs.30Other studies also demonstrated that endovascular cooling achieved a significant reduction in core body temperature.31External application of ice was used in the present study to induce hypothermia because this method is easy to implement. Even if the trauma team is outside the hospital, as long as there is an ice bag, the temperature-decreasing process can be done. If inside a hospital, intravenous infusion of cold saline can be selected to better control the temperature-decreasing process.

Hemorrhagic coagulopathy is a significant complication after traumatic injury, which has an important impact on the development and therapy of traumatic-hemorrhagic shock.3,32–33Studies have shown that hypothermia has a very important effect on coagulation if the core temperature is less than 34°C,34,35whereas hypothermia more than 34°C has no or little effect on coagulation. Hence, 34°C of hypothermia was selected in the present study. Resnick36reported that hypothermia more than 34°C did not affect the coagulation in hemorrhagic-shock pigs; all coagulation parameters, including TT, APTT, platelet count, activated clotting time, and thromboelastogram did not differ between that at 34°C and at 37°C. To observe the influence of the resuscitation regimen in the present study on coagulation function in hemorrhagic shock, we observed changes of coagulation parameters such as TT, PT, PT-INR, APTT, fibrinogen, platelet count, and platelet aggregation. The results showed that after hemorrhagic shock, coagulation functions had different levels of impairment: it appeared that APTT was significantly prolonged. During hypothermia and hypotensive resuscitation, TT, PT, PT-INR and APTT were further prolonged or increased as compared with the end of phase I. However, among these parameters, PT, PT-INR, and APTT in the hypothermia (34°C) group were slightly higher than in the normothermia (37°C) group, and platelet aggregation in the hypothermia group was slightly lower than in the normothermia group at the same pressure level. These results showed that severe hemorrhage impairs the coagulation function, and that mild hypothermia applied for a limited period before bleeding only slightly inhibited coagulation function.

The present study showed that permissive hypotensive resuscitation in combination with hypothermia had a beneficial effect upon uncontrolled hemorrhagic shock, and that short-term and mild hypothermia could increase the beneficial effect of hypotensive resuscitation. Nevertheless, the present study had limitations. First, only small animals were used; whether this effect can be extrapolated to larger animals and humans needs further investigation. Second, the duration of hypotensive and hypothermic resuscitation during uncontrolled hemorrhagic shock was only 1 h; discovering the optimal duration needs further study.

Permissive hypotensive resuscitation in combination with mild hypothermia had a synergistic effect upon uncontrolled hemorrhagic shock in rats. Short-term and mild hypothermia could increase the beneficial effect of hypotensive resuscitation. The latter could reduce blood loss and stabilize the concentration of hemoglobin. Mild hypothermia could protect mitochondrial function and decrease oxygen consumption and energy metabolism, but did not cause severe damage to coagulation function. This combination had a synergistic effect, including protection of organ function, hemodynamic stabilization, and animal survival.

The authors thank Tang Shihai, M.D., Associate Professor in Statistics in the editorial office of the Journal of Trauma Surgery, Chongqing, China, for assistance in the statistical analysis in the preparation of this manuscript.