Resuscitation 58 (2003) 103 /112 www.elsevier.com/locate/resuscitation
Rapid rewarming after mild hypothermia accentuates the inflammatory response after acute volume controlled haemorrhage in spontaneously breathing rats Per Vaagenes *, Yngvar Gundersen, Per Kristian Opstad Norwegian Defence Research Establishment, Division of Protection and Material, N-2027 Kjeller, Norway Received 18 September 2002; received in revised form 2 December 2002; accepted 20 February 2003
Abstract Accidental hypothermia is a common companion of trauma/haemorrhage, and several clinical studies have identified reduced body temperature as an independent risk predisposing to increased morbidity and mortality. Accordingly, the majority of trauma care guidelines prescribe early and aggressive rewarming of hypothermic patients. Enzyme reactions are generally downregulated at temperatures below 37 8C, including most of those responsible for the inflammatory response. The rationale for adhering to these recommendations uncritically may therefore be questioned. In a rat model of mild hypothermia and haemorrhagic shock we wanted to compare the influence of rapid rewarming with persistently reduced temperature on the synthesis of early inflammatory mediators and organ function. Thirty-four male albino Sprague /Dawley rats were studied. Withdrawal of 2.5 ml blood/100 g body weight was performed over 10 min, with simultaneous reduction of body temperature to 32.5 /33.5 8C. Seventy-five minutes after initiation of bleeding, two-thirds of the shed blood was retransfused. One group (n /17) was rewarmed to normothermia, the other (n /17) was kept hypothermic. The study was terminated after an observation period of 2 h. At the end of the study the rewarmed animals had a significantly lower mean arterial pressure, higher heart rate, higher synthesis of reactive oxygen species from peritoneal phagocytes, increased circulating levels of nitric oxide, and higher values of the organ markers aspartate aminotransferase and urea. The proinflammatory cytokines TNF-a and IL-6, the anti-inflammatory cytokine IL-10, the organ markers alanine aminotransferase, aglutathione S -transferase and creatinine, as well as organ injury scores were equal in both groups. Three rewarmed rats died prematurely, versus one hypothermic animal. In conclusion, the results suggest that during the early stages after haemorrhagic shock, rapid rewarming from mild hypothermia may have unfavourable effects both on basic haemodynamic variables, and on the internal inflammatory environment of cells and tissues. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Haemorrhage; Shock; Resuscitation; Hypothermia; Inflammatory response
Resumo A hipotermia acidental acompanha com frequeˆncia o trauma/hemorragia, e va´rios estudos clı´nicos identificaram a baixa temperatura corporal como um risco independente predispondo ao aumento da mobilidade e da mortalidade. De acordo com isto, a maioria das recomendac¸o˜es de tratamento do trauma prescrevem o reaquecimento agressivo e precoce dos doentes hipote´rmicos. As reacc¸o˜es enzima´ticas sa˜o geralmente reguladas no sentido descendente com temperaturas inferiores a 37 8C, incluindo a maioria daquelas responsa´veis pela resposta inflamato´ria. A raza˜o para aderir de forma na˜o crı´tica a estas recomendac¸o˜es pode, portanto, ser questionada. Num modelo de hipotermia ligeira e choque hemorra´gico do ratinho quisemos comparar a influeˆncia do reaquecimento ra´pido com a manutenc¸a˜o persistente de temperatura reduzida, na sı´ntese de mediadores precoces da inflamac¸a˜o e a func¸a˜o dos orga˜os. Foram estudados 34 ratinhos albinos machos Sprague-Dawley. Foram retirados 2.5 ml de sangue/100 g de peso, durante 10 min, com reduc¸a˜o simultaˆnea da temperatura corporal para 32.5 /33.5 8C. Setenta e cinco minutos depois do inicio da hemorragia, foi retransfundido 2/3 do sangue retirado. Um grupo (n/17) foi reaquecido ate´ normotermia, o outro (n /17) foi
* Corresponding author. E-mail addresses: [email protected]
(P. Vaagenes), [email protected]
(Y. Gundersen), [email protected]
(P.K. Opstad). 0300-9572/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0300-9572(03)00102-3
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
mantido hipote´rmico. O estudo terminou apo´s 2 h de observac¸a˜o. No final do estudo, os animais reaquecidos tinham presso˜es arteriais me´dias significativamente menores, frequeˆncias cardı´acas maiores, maior sı´ntese pelos fago´citos peritoneais de radicais reactivos ao oxige´nio, nı´veis mais elevados de o´xido nı´trico circulante e nı´veis maiores dos marcadores de lesa˜o de orga˜o aspartatoaminotransferase e ureia. As citoquinas pro´-inflamato´rias TNF-a e IL-6, a citoquina anti-inflamato´ria IL-10, os marcadores de lesa˜o de orga˜o alanina-aminotransferase, a-glutationa S-tranferase e creatinina, bem como os ´ındices de lesa˜o de orga˜o, foram iguais em ambos os grupos. Treˆs ratinhos reaquecidos morreram prematuramente, contra um dos hipote´rmicos. Em conclusa˜o, os resultados sugerem que o reaquecimento da hipotermia ligeira, durante os estados precoces do choque hemorra´gico, pode ter efeitos desfavora´veis nas varia´veis hemodinaˆmicas e no ambiente inflamato´rio interno das ce´lulas e dos tecidos. # 2003 Elsevier Ireland Ltd. All rights reserved. Palavras chave: Hemorragia; Choque; Ressuscitac¸a˜o; Hipotermia; Resposta inflamato´ria
Resumen La hipotermia accidental es una acompan˜ante frecuente de trauma/hemorragia, y varios estudios clı´nicos han identificado temperatura corporal reducida como factor de riesgo independiente que predispone a mayor morbilidad y mortalidad. Por ello, la mayorı´a de las guı´as de cuidado de trauma indican recalentamiento precoz y agresivo de pacientes hipote´rmicos. Las reacciones enzima´ticas esta´n generalmente subreguladas a temperaturas bajo 37 8C, incluyendo la mayorı´a de aquellas responsables de la respuesta inflamatoria. La razo´n para adherir estas recomendaciones sin crı´tica puede ser cuestionada. Quisimos comparar la influencia que tendrı´a el recalentamiento ra´pido y la persistencia de temperatura reducida, sobre la sı´ntesis de mediadores tempranos de la inflamacio´n y en la funcio´n orga´nica, en un modelo de hipotermia moderada y shock hemorra´gico en rata. Se estudiaron 34 ratas macho Sprague-Dawley albino. Se les retiro´ 2.5 ml de sangre/ 100g de peso corporal en 10 minutos, con reduccio´n simulta´nea de la temperatura corporal a 32.5 /33.5 8C. Setenta y cinco minutos despue´s de el inicio del sangrado, dos tercios de la sangre extraı´da fue retransfundida. Un grupo (n/ 17) fue recalentado hasta normotermia, el otro (n /17) fue mantenido hipote´rmico. El estudio fue terminado despue´s de un perı´odo de observacio´n de 2 h. Al final del estudio los animales recalentados tenı´an una presio´n arterial media significativamente menor, mayor frecuencia cardı´aca, mayor sı´ntesis de especies reactivas al oxı´geno de los fagocitos peritoneales, niveles circulantes de oxido nı´trico aumentado,, y mayores valores de marcadores de organos aspartato aminotransferasa y urea. Las citoquinas proinflamatorias TNF-a y IL-6, la citoquina inflamatoria IL-10, los marcadores de organos alanino aminotransferasa, a-glutatio´n, S -transferasa y creatinina, al igual que los puntajes de lesiones, fueron iguales en ambos grupos. Tres ratas recalentadas murieron prematuramente, versus un animal hipote´rmico. En conclusio´n, los resultados sugieren que durante las etapas tempranas despue´s de shock hemorra´gico, el calentamiento ra´pido de hipotermia moderada puede tener efectos desfavorables tanto en las variables hemodina´micas ba´sicas, como sobre el ambiente inflamatorio interno de ce´lulas y tejidos. # 2003 Elsevier Ireland Ltd. All rights reserved. Palabras clave: Hemorragia; Shock; Reanimacio´n; Hipotermia; Respuesta inflamatoria
1. Introduction Accidental hypothermia is often encountered during the management of severely injured patients, and clinical studies have documented a relationship between hypothermia, morbidity and mortality [1 /5]. The severity of injury may lead to immediate demise at the scene of the accident, while uncontrollable haemorrhage is a common cause of death both during transport and up to several hours after arrival in hospital . Haemorrhage results from mechanical damage or coagulopathy. The former may be surgically repaired, while coagulopathy is a multifactorial and insidious process primarily caused by the effects of trauma and the immediate resuscitative measures. Reduced body temperature and excessive activation of the inflammatory cascade systems may further aggravate the bleeding diathesis [6 /9]. In addition, the adverse influence of massive transfusions on coagulation is well-documented .
Hypothermia is commonly considered a major cause of life-threatening posttraumatic complications. The precise mechanisms accounting for this ominous impact are not clearly understood, and a discrepancy exists between clinical experience and experimental studies [11 /13]. Current trauma care guidelines call, almost unanimously, for active efforts at preventing hypothermia. Once it has occurred, aggressive treatment is recommended . In a volume controlled haemorrhagic shock model with spontaneously breathing rats, we have earlier shown that lowering core temperature to 32.5 /33 8C during the shock period, and retaining a low temperature during the resuscitation and post-haemorrhagic phase, decreased the severity of the early inflammatory responses and preserved short term organ function better than persistent normothermia . Based on the prevailing trauma care guidelines we wanted to use the same hypothermic shock model and investigate how the proximal inflammatory responses and outcome are influenced by early and rapid rewarming to normother-
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
mia during the resuscitation phase. When designing the model, we tried to mimic the clinical situation as far as possible.
2. Materials and methods 2.1. Animals and surgical procedure The study was performed according to the recommendations issued by the European Council for Laboratory Animal Science and after approval of the Local Ethics Committee. Thirty-four male albino Sprague/Dawley rats were used. They had free access to food and water. The experiments started at 07:30 a.m. Anaesthesia was induced with a mixture of nitrous oxide, oxygen, and isoflurane in a semiclosed container. Induction time was less than 2 min for all animals. During the rest of the study the rats breathed spontaneously the gas mixture via a facemask. The anaesthetic depth was regulated to a level just sufficient to avoid reaction to tail pinching. For monitoring of heart rate (HR) and core temperature, ECG (4-lead) electrodes were attached to the paws, and temperature probes (thermistors) placed in the rectum and oesophagus. With supplemental local anaesthesia a sterile cut-down in the groin was performed. The femoral artery was isolated and cannulated with a 0.8 mm cannula (Venflon, Viggo, Denmark) for blood withdrawal, volume replacement, and continuous recording of arterial pressure. Preparation time was less than 30 min for all rats. 2.2. Experimental procedure To avoid day-to-day variations the experiments and measurements were done in pairs, one from each group. During preparation core temperature was controlled at 37.5 /38 8C, defined as normothermia . Room temperature was kept at 23 8C. Baseline recordings were done after a control period of 10 min. Thereafter a steady withdrawal of 2.5 ml of blood/100 g body weight from the arterial cannula was undertaken over 10 min. The fall in mean arterial pressure (MAP) was registered intermittently. The blood was collected in small heparinised syringes (Heparin 7.5 IU/ml). Cooling was started simultaneously with initiation of bleeding. For this purpose ethyl alcohol (70%) was sprayed on shaved abdomen until core temperature had fallen to 32.5 / 33.5 8C . The shock period lasted for 75 min. The rats were randomised either to be rewarmed to 37 8C during resuscitation, or to remain hypothermic throughout the experimental period. External heating was supplied by an infrared lamp . At 75 min two-thirds of the shed blood was transfused over 10 min. To maintain open arterial lines and compensate for blood sampling, an infusion of 1 ml/100 g body weight per
Fig. 1. Protocol. Haemorrhagic shock was induced by withdrawal of 2.5 ml blood/100 g body weight over 10 min. Simultaneously cooling to 339/0.5 8C was started. Seventy-five minutes after initiation of bleeding, all animals were retransfused with two-thirds of the collected blood. One group of animals (n/17) were rewarmed to normothermia, the others (n/17) were kept hypothermic. Both groups were thereafter observed for another 2 h.
hour of normal saline was established. After resuscitation the animals were observed under light anaesthesia for another 2 h, giving a total insult time of 195 min (Fig. 1). They were sacrificed by exsanguination. The heart, lungs, liver, small intestines, and kidneys were immediately removed and examined for pathological changes, including discoloration, bleeding, consolidation (atelectasis in lungs), ischaemic changes and oedema. Based on a scale from 0 to 3, where zero represents no observed pathological changes, one slight changes, two moderate changes and three severe changes, each animal was given a pathological score ranging from 0 (normal) to 15 (severe changes in all five organs). The scoring was done unblinded, but independently, by two examiners.
2.3. Monitoring of vital functions HR and systolic, diastolic, and MAPs were continuously monitored (Grass Polygraph Model 7E, Grass Instrument Co, Mass., USA), as was rectal (Tr) and oesophageal (Tc) temperatures (Cormak K 90001 thermometer, UK and Tes-1303 digital thermometer, Tess Electrical Electronic Corp., Taiwan). Respiratory rate (RR) was counted at baseline, 15, and 75 min after initiation of haemorrhage, and at 15 and 120 min after start of resuscitation.
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
2.4. Blood sampling and assays Blood gases and acid base state were measured at the same time points (ABL 330, Acid-Base Laboratory, Radiometer, Copenhagen, Denmark). Total white blood cells (WBC), granulocytes, lymphocytes and platelets were counted (Unipath Cell-Dyn 610 Automatic counter, 850 Maude Avenue, Mountain View, CA) at baseline and at time points 75 and 195 min. Plasma values of the pro-inflammatory cytokines TNF-a and IL-6, the anti-inflammatory substances IL-10 and corticosterone, and markers of organ function and integrity were determined at the end of the study (alanine aminotransferase (ALT), aspartate aminotransferase (AST), and aglutathione S -transferase (a-GST) for hepatocellular damage, and creatinine/urea as measures of renal function). To quantitate nitric oxide (NO) we measured its oxidative product nitrite with a colorimetric assay, after reaction with the Griess reagent.
recordings were done for 60 min, and the integral of the response curve was calculated. 2.7. Statistical analysis Data are presented as mean9/S.D. Differences between groups were tested with paired t-test, or Wilcoxon signed rank sum test when appropriate. For differences within groups ANOVA for repeated measurements followed by Student-Newman-Keul’s post hoc test was applied. A regression model was fit with pathological score as outcome, and univariate analysis performed in order to isolate independent predictors among haemodynamic and various laboratory parameters. For continuous variables a log transformation was used. A P value equal or less than 0.05 was considered statistically significant.
3. Results 2.5. Peritoneal cells At the end of each experiment a mini-laparotomy was performed and peritoneal cells obtained by washing the peritoneum with 35 ml 0.9% NaCl. The peritoneal washout was observed for cloudiness and colour as a reflection of haemolysis respectively cell content, and samples were examined by light microscopy. Peritoneal cells were centrifuged, washed and resuspended in Hanks balanced salt solution supplied with 20 mmol/l Hepes buffer and 5 mmol/l glucose. Cell numbers were counted in a 610 Hematology Analyzer (Seqoia Turner Corp., Mountain View, CA). To induce hypotonic lysis of contaminating erythrocytes, the cell suspension was incubated with 0.83% NH4Cl, centrifuged and washed in 0.9% NaCl. 2.6. Chemiluminescence The respiratory burst in the peritoneal cells was activated by the protein kinase C activator phorbol myristate acetate (PMA) or the surface receptor agonist N -L-formyl-L-methionyl-L-leucyl-L-phenylalanine (fMLP). Luminol-(5-amino-2,3-dihydro-1,4-phtalazine dione) and lucigenin-(N ,N ?-dimethyl-9,9?bisacridinium dinitrate) (Sigma Chemical, St Louis, MO) enhanced chemiluminescence (CL) responses were measured in a Labsystems Luminoskan luminometer (Turku, Finland) with samples in 96 well disks. The collected peritoneal cells (3 /105 cells per well) were exposed to 107 M PMA and 10 4 M luminol, 10 7 M lucigenin, 106 M fMLP and 10 7 M luminol, and 106 M fMLP and 107 M lucigenin in each well. Luminol-enhanced CL was also measured after addition of horseradish peroxidase (HRP), 1 U HRP/well. The
Seventeen parallel experiments were done in 34 animals. There were no significant differences in physiologic variables between the groups either at baseline or at the end of the shock phase (Table 1). 3.1. Survival Three rats in the rewarmed group became apnoeic with bradycardia and hypotension, and died between time points 160 and 180 min, in spite of resuscitative efforts. One hypothermic rat died at the end of study (195 min). Early after shock these rats did not differ compared to their counterparts, either with regard to arterial pressure, RR or base deficit. Thus, no indication of cardiopulmonary or metabolic failure was seen at this stage. However, immediately before death a severe base deficit had developed, and there was an abrupt increase in serum potassium. 3.2. Haemodynamic parameters Haemorrhage lowered MAP significantly and to the same extent in both groups (Fig. 2), but at the end of the study a significantly (P B/0.01) higher MAP was measured in the hypothermic animals (1219/14 mmHg vs. 959/20 mmHg). The hypothermic rats also had less tachycardia (3279/59 vs. 3749/45 beats/min; P B/0.01) and had a significantly lower RR (579/15 vs. 759/14, P B/0.01). 3.3. Acid/base, serum electrolytes and other physiologic variables Within 15 min from the onset of bleeding base excess (BE) fell abruptly and in parallel in both groups from
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
Table 1 Physiologic variables at baseline and end of shock phase (mean9/S.D.)
Body weight (g) Hct (%) Rectal temperature (8C) MAP (mmHg) HR (beats per minute) RR (per minute) p O2 (kPa) p CO2 (kPa) pH BE (base excess)
Rewarmed group (n/17) (baseline)
Hypothermic group (n/17) (baseline)
Rewarmed group (n /17) (end of shock phase)
Hypothermic group (n/17) (end of shock phase)
3429/61 40.89/1.6 37.29/0.5 1029/10 3599/47 739/12 17.99/3.5 6.39/1.3 7.379/0.07 0.39/1.6
3479/72 41.19/1.8 37.29/0.6 1099/13 3919/61 739/17 19.49/4.4 6.39/1.5 7.379/0.05 1.19/1.8
32.19/1.7 32.89/0.4 1039/15 3029/38 639/11 23.49/3.8 5.99/1.8 7.219/0.07 /10.69/5.0
32.19/1.8 32.89/0.3 1099/16 3089/44 649/16 26.59/3.9 5.89/1.6 7.209/0.07 /12.29/4.8
0.39/1.6 to /8.89/3.2 in the group to be rewarmed and from 1.19/1.8 to /8.09/2.9 in the hypothermic group, and continued to fall during the shock phase (Table 1). A slight improvement was thereafter observed in the surviving rats. There were no differences between the groups with regard to arterial pH, p CO2 or pO2 changes in the surviving animals during the study, neither with nor without correction for body temperature. Blood lactate was significantly increased to 8.19/2.3 vs. 8.99/ 3.1 mmol/l (P /0.76), in the rewarmed and hypothermic group respectively at the end of the study. Serum Na, K , and Ca remained within normal limits in surviving animals without a difference between groups. The number of platelets declined insignificantly during the study period in both groups. As expected, serum glucose was increased and tended to be higher in the hypothermic group towards the end of the study (16.19/6.8 vs. 13.29/4.9 mmol/l, (P /0.14).
3.4. Pro-and anti-inflammatory mediators Total WBC count dropped significantly as a result of haemorrhage in both groups, but returned to baseline values at the end of the study. Differential count, however, showed that the granulocyte population had increased throughout the study period from 1.69/0.9 / 109/l to 2.89/0.9 /109/l, P B/0.02, and from 1.49/0.4 / 109/l to 3.29/1.8 /109/l, P B/0.005, in the rewarmed and hypothermic group respectively. Microscopy of the peritoneal washout showed that animals with the worst intestinal damage had a high content of polymorphonuclear neutrophils (PMNs), erythrocytes, and of living micro-organisms. In animals without damage the fluid was virtually clear. At measure point 90 min the proximal pro-inflammatory cytokine TNF-a was elevated in both groups to 5559/732 pg/ml in the rewarmed vs. 2449/230 pg/ml in the hypothermic rats (P / 0.12), (Fig. 3). At the end of the study a decline to 1569/313 pg/ml was measured in the former with virtually no change in the latter group (2639/116 pg/ml). IL-6 was only determined at the end and did not reveal any difference between groups (56319/11343 pg/ml in the rewarmed rats vs. 31069/ 3892 in the hypothermic, P / 0.42). The production of the anti-inflammatory cytokine IL-10 tended to be lower at time point 195 min in the rewarmed group (3339/225 vs. 8109/806 pg/ml, P / 0.08). Compared to the normal value of 3979/4 nmol/l established at our laboratory, the plasma value of the endogenously produced anti-inflammatory substance corticosterone increased significantly in the hypothermic group (12169/265 nmol/l, P B/0.0005). Also in the rewarmed group an increase was measured (8909/317 nmol/l) but less marked (P / 0.03 between groups). 3.5. Production of free radicals
Fig. 2. Development of MAP (mean9/S.E.M.). *P B/0.05 vs. preceding measurement; #P B/0.05 between groups.
The synthesis of oxygen free radicals was measured by CL in peritoneal cells from all animals. As may be seen
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
in Fig. 4, oxygen radical production was significantly reduced in the hypothermic group, and P -values less than 0.05 were observed with combinations of luminol/ fMLP and lucigenin/PMA. The same pattern of changes was obtained after measurement of NO in serum (29.89/ 4.4 mM in the rewarmed vs. 24.89/2.3 mM in the hypothermic group, P B/0.01). 3.6. Markers of cell integrity and function Plasma AST, ALT and a-GST were used as markers of liver function. The results in the rewarmed and the hypothermic group were as follows: AST 4939/89 vs. 2549/45 IU/l (P / 0.03), ALT 7639/280 vs. 9679/951 IU/l (P /0.66), and a-GST 123969/13160 vs. 165639/ 33854 mg/l (P / 0.80). To assess kidney function plasma creatinine and urea were measured. The corresponding findings for the two groups were: creatinine 1269/78 vs. 1369/82 mmol/l (P / 0.76), and urea 9.29/1.8 vs. 7.89/ 1.5 mmol/l (P/ 0.02). 3.7. Organ damage The scores of the individual organs are shown in Table 2. Except for the intestine, scores were numerically lower for all organs in the hypothermic group, but due to the large standard deviation mean values were not significantly different. When adding up the scores from the individual organs in all animals, mean score was 3.779/2.77 vs. 2.929/2.18 (P =0.3) in the rewarmed and hypothermic group respectively. Regression analysis showed that there was a significant relationship between bad score and elevated a-GST, AST, IL-6,
Fig. 4. CL in peritoneal phagocytes after stimulation with the protein kinase C activator PMA or the surface receptor agonist fMLP, (mean9/S.E.M.). Luminol and lucigenin enhanced CLs were measured in a luminometer. #P B/0.05 between groups.
TNF-a, IL-10, creatinine and arterial base deficit at time point 195 (Table 3), but not with ALT, urea, and blood lactate. At time point 90 there was no relationship between score and TNF-a.
4. Discussion Compared to prolonged, mild hypothermia, rapid rewarming during resuscitation in this rat model of controlled haemorrhagic shock did not attenuate early inflammatory responses positively nor the production of reactive oxygen species (ROS) or NO. Neither did it influence early organ damage positively nor increase the ability to survive the first phases of shock. Rewarming during resuscitation from induced, mild hypothermia however, may be less injurious than if normothermia is maintained both during and after haemorrhage [14 /16]. Table 2 Pathological scores of individual organs in the two groups at gross autopsy examination (mean9/S.D.) Rewarmed group (score) Hypothermic group (score)
Fig. 3. Plasma values of TNF-a 75 min after initiation of bleeding and at the end of the study (mean9/S.E.M.). No significant difference was seen between groups.
Heart Lungs Liver Small intestine Kidneys
0.339/0.36 0.389/0.65 1.549/1.08 1.389/0.58 0.239/0.40
NS between groups.
0.229/0.23 0.159/0.32 1.239/1.12 1.389/0.76 0.159/0.36
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
In our study rewarming to above 37 8C was achieved within 60 min. Rapid rewarming is frequently used in small animal research [13,15,17], and although slower rates are most commonly used clinically [6,11], rapid rewarming may also be achieved with special techniques . Generally rapid shifts of internal environment of body cells cannot be recommended, and several investigators have observed detrimental consequences from posthypothermic rewarming [18,19]. In an effort to simulate a ‘pre-hospital clinical shock scenario’, we allowed blood pressure to drop spontaneously, and early recovery was determined consequently by the body’s natural physiological homeostatic response. In the field this means compensatory cardiovascular, respiratory and metabolic measures until further treatment is available. Resuscitation included only partial re-transfusion without supplementary intravenous hydration except for compensation for blood sampling, thus extending the insult period to 195 min. In concert with other haemorrhagic shock studies [13,20], arterial pressure was better preserved in the hypothermic animals (Fig. 2). There was, however, no correlation between organ damage (pathological score) and MAP at any time period, indicating that arterial pressure may not be a single, determinant outcome factor [20,21]. Indeed Mizushima et al.  showed that after trauma /haemorrhage in rats, rewarming during resuscitation improved cardiac output and stroke volume compared to maintaining normothermia during haemorrhage or maintenance of prolonged hypothermia after resuscitation, while MAP did not differ between groups. The acute inflammatory response is composed of an elaborate cascade of both pro- and anti-inflammatory mediators, and the balance between these mediators is an important determinant for the outcome after injury. A key factor at an early stage is the activation of PMNs and their interaction with endothelial cells at the site of injury . Originally this is a local response vital for a successful outcome and aimed at preventing a systemic reaction which may be deleterious to remote tissues and organs. However, the activation after major trauma may Table 3 Regression analysis showing the relationship between pathological score and selected parameters in the two groups at the end of the study (mean9/S.E.M.)
a-GST AST IL-6 TNF-a IL-10 Creatinine BE
0.70 0.50 0.54 0.59 0.44 0.45 0.59
B/0.0001 B/0.005 B/0.01 B/0.01 B/0.02 B/0.02 B/0.005
28 28 28 28 26 28 26
sometimes be so strong that an overwhelming systemic response is initiated. The resulting overproduction of pro-inflammatory cytokines promotes complement activation, further increases leukocyte /endothelial adhesion, endothelial permeability, coagulopathy, and activation of the hypothalamic /pituitary /adrenocortical (HPA) axis. Subsequently the clinical picture of the traumatised patient becomes dominated by systemic hypotension, decreased organ perfusion, cellular energy deficit, capillary leakage, and multiple organ dysfunction or failure. We observed a significant increase of PMN cells during the study period, and these cells also dominated in the peritoneal washout of the animals with the worst intestinal damage irrespective of group. This indicates that the haemorrhagic insult initiated both a systemic and local recruitment of PMN cells as a reflection of an early pro-inflammatory response. Severely enhanced circulating levels of the proximal cytokines TNF-a and IL-6 have been demonstrated both experimentally and clinically after haemorrhagic shock , and reducing their synthesis has been shown to influence organ injury and mortality positively . In the same haemorrhagic shock model we have previously demonstrated that, compared to persistent normothermia, moderate reduction of body temperature to 32.5 /33 8C significantly decreased IL-6 and tended to decrease TNF-a and the anti-inflammatory cytokine IL10, while corticosterone increased . In the present study TNF-a reached higher levels at time point 90 min in the rewarmed group, but then decreased towards the end of the study while the lower TNF-a levels in the hypothermic animals did not change. At time point 195 min IL-6 was elevated, and somewhat more in the rewarmed animals. But due to great variability in cytokine levels the difference was not significant. IL-10 tended to be higher at time point 195 min in the hypothermic animals (P /0.08). This is in contrast with previous studies where IL-10 tended to be lower in the hypothermic animals . Variability in cytokine levels, however, has been registered by others, and without a plausible explanation . Factors related to animal model used, and type of insult, site and time of neutrophil harvest, and analytical methods are probably important. Corticosterone was significantly elevated in both groups but highest (P /0.03) in the hypothermic animals. It is known that both hypothermia and trauma may activate the HPA axis and thus stimulate glucocorticoid secretion which is a strong inducer of IL-10 . In the present study persistent low body temperature seemed to maintain plasma corticosterone increase and tended to promote an anti-inflammatory IL-10 increase. The adverse effects from an abundance of inflammatory substances may thus be ameliorated. Other investigators have shown that although proinflammatory forces may prevail at local sites of injury (or infection), the same cytokines often do not dominate
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
in the systemic circulation . This may also hold true for our study in which severe pathological changes were identified in ileum and caecum. The results may therefore indicate that an aggressive rewarming process could influence the pro- and anti-inflammatory responses in an unfavourable way. Based on gross autopsy studies of selected organs, pathological scoring systems are used to assess organ damage . In the present study blinding unfortunately was not accomplished, but a thorough evaluation was undertaken by two independent observers, and their scoring was practically identical. In line with other investigations we found that ileum and caecum were most heavily damaged . The finding of a significant relationship between gross pathological damage of selected organs on one hand, and the observed increase of base deficit, certain pro-inflammatory substances and indicators of hepatic or renal dysfunction on the other, should justify the use of our model for early outcome studies after haemorrhagic shock in rats. There was a highly significant correlation between pathological score and a-GST, and also with AST, but not with ALT. a-GST is supposed to be an early and reliable marker of liver damage whereas ALT may be temporarily prevented in entering the systemic circulation due to disturbances in the hepatic microcirculation. The increase of AST probably partly reflects general organ damage and was as in previous studies lowest in the hypothermic group . There was also a significant correlation between pathological score and arterial base deficit, but not with blood lactate. Acid base values are commonly used as indicators of abdominal injury  and blood lactate as an index of tissue oxygenation . Lactate clearance is strongly dependent on liver function, whose function may suffer at an early stage of haemorrhagic shock. Our study seemed to show that the ileum was the first organ to reveal pathological changes, and the identification of cellular elements in the peritoneal washout, especially leukocytes and macrophages, red blood cells, and microorganisms strongly indicated translocation from inflamed intestinal tissue. In this respect the results suggest that induced hypothermia followed by early and fast rewarming in the resuscitation phase is better than persistent normothermia, but less beneficial than persistent, moderate hypothermia. It should be emphasized, however, that both timing and duration of rewarming conceivably are important factors . The same applies to the rate of rewarming, which was relatively aggressive. In less than 60 min body temperature was raised from 32.5 to 37 8C. The inhibitory effect of low temperatures on enzymes of the coagulation system has been of special concern . Our results tend to suggest less rigid conclusions. Certainly, using a model of controlled haemorrhage, there was no need to care about ongoing blood loss. But also in many clinical situations bleeding is a minor
problem. Under such circumstances more attention may be paid to the ability of low temperatures to curb reactions responsible for an early exuberant post-injury pro-inflammatory response. Reduced high-energy phosphate utilisation and induction of a protective heat response may also be attained . Taken together, this may more than outweigh the detrimental effects upon coagulation, especially when speaking about moderate hypothermia, which has for many years been used successfully during planned surgical interventions. The synthesis of NO was significantly reduced in the hypothermic rats compared to their rewarmed counterparts. The overall consequences for organ function and integrity may be difficult to assess. The continuous release of NO from the constitutive endothelial isoform of NO synthase (ecNOS) maintains the vasculature in a state of active vasodilatation and inhibits the adhesion of platelets and granulocytes to the endothelial surface. Initially, haemorrhagic shock is known to reduce the production of NO from ecNOS . Hypothermia may amplify this effect . Theoretically maldistribution of the remaining cardiac output may ensue and the microcirculation suffer . However, as the inducible NO synthase becomes expressed, the early state of nitrosopenia is gradually replaced by NO overproduction, clinically reflected in a global vasodilatation and reduced blood pressure. At the end of the study MAP was significantly better preserved in the hypothermic animals. Hypothermia also significantly reduced the production of ROS in peritoneal cells. In biological systems superoxide (O2+ ) rapidly combines with NO to form the potent noxious substance peroxynitrite (ONOO ), which is probably the prime cytotoxic species in many inflammatory conditions . Both absolute and relative fluxes of NO and O2+ are important for the formation of peroxynitrite, and thus also for the resultant tissue injury . Anti-peroxinitrite strategies have been investigated in several experimental animal shock models to reduce oxidant tissue injury [38,39]. The positive effects of hypothermia seen in this study may partly work via the same mechanisms. As discussed earlier, the model has several limitations, including the absence of a true ‘accidental’ hypothermia, no ongoing bleeding, and, maybe most importantly, a relatively short observation period. Also anaesthesia and the use of heparin as an anticoagulant may influence our measurements. However, in any case all these factors would have the same effects in the two groups and are also minimised by the pairwise study design. In conclusion we did not observe overall gains of rapid rewarming in our short term volume controlled haemorrhagic shock model. In concert with other studies , prolonged, mild hypothermia tended to maintain greater haemodynamic stability and protect the hypothermic animals. In defined clinical situations
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
refraining from aggressive rewarming may constitute a way to dampen hyperactivation (and subsequent exhaustion) of important immune functions, and thus contribute to a reduced incidence of post-traumatic complications.
We thank Professor Torstein Lyberg and Trude Asplin at Ulleva˚l University Hospital, Oslo, for valuable help and advice and for measuring the nitric oxide levels in serum.
References  Luna GK, Maier RV, Pavlin EG, Anardi D, Copass MK, Oreskovich MR. Incidence and effect of hypothermia in seriously injured patients. J Trauma 1987;27:1014 /8.  Steinemann S, Shackford SR, Davis JW. Implication of admission hypothermia in trauma patients. J Trauma 1990;30:200 /2.  Jurkovich GJ, Greiser WB, Luterman A, Curreri PW. Hypothermia in trauma victims: an ominous predictor of survival. J Trauma 1987;27:1019 /24.  Gentilello LM, Jurkovich GJ, Stark MS, Hassantash SA, O’Keefe GE. Is hypothermia in the victim of major trauma protective or harmful? A randomized, prospective study. Ann Surg 1997;226:439 /49.  Seekamp A, Ziegler M, van Grinsveen M, Grotz M, Regel G. The role of hypothermia in trauma patients. Eur J Emerg Med 1995;2:28 /32.  Gentilello LM, Pierson DJ. Trauma critical care. Am J Respir Crit Care Med 2001;163:604 /7.  Watts DD, Trask A, Soeken K, Perdue P, Dols S, Kaufmann C. Hypothermic coagulopathy in trauma: effect of varying levels of hypothermia on enzyme speed, platelet function, and fibrinolytic activity. J Trauma 1998;44:846 /54.  Johnston CD, Chen Y, Reed RI. Relative sensitivity of the clotting cascade to hypothermia. Surg Forum 1989;40:199 /201.  Smith CE, Yamat RA. Avoiding hypothermia in the trauma patient. Curr Opin Anaesthesiol 2000;13:167 /74.  Cosgriff N, Moore EE, Sauaia A, Kenny-Moynihan M, Burch JM, Galloway B. Predicting life-threatening coagulopathy in the massively transfused trauma patient: hypothermia and acidosis revisited. J Trauma 1997;42:857 /61.  Mikhail J. The trauma triad of death: hypothermia, acidosis, and coagulopathy. AACN Clin Iss 1999;10:85 /94.  Meyer DM, Horton JW. Effect of moderate hypothermia in the treatment of canine haemorrhagic shock. Ann Surg 1988;207:462 /9.  Takasu A, Carrillo P, Stezoski W, Safar P, Tisherman SA. Mild or moderate hypothermia but not increased oxygen breathing prolongs survival during uncontrolled haemorrhagic shock in rats, with monitoring of visceral dysoxia. Crit Care Med 1999;27:1557 /63.  Gundersen Y, Vaagenes P, Pharo A, Valø ET, Opstad PK. Moderate hypothermia blunts the inflammatory response and reduces organ injury after acute haemorrhage. Acta Anaesthesiol Scand 2001;45:994 /1001.  Mizushima Y, Wang P, Cioffi WG, Bland KI, Chaudry IH. Should normothermia be restored and maintained during
resuscitation after trauma and hemorrhage? J Trauma 2000;48:58 /65. Mizushima Y, Wang P, Cioffi WG, Bland KI, Chaudry IH. Restoration of body temperature to normothermia during resuscitation following trauma /haemorrhage improves the depressed cardiovascular and hepatocellular functions. Arch Surg 2000;135:175 /81. Prueckner S, Safar P, Kentner R, Stezoski J, Tisherman SA. Mild hypothermia increases survival from severe pressure-controlled hemorrhagic shock in rats. J Trauma 2001;48:58 /65. Suehiro E, Povlishock JT. Exacerbation of traumatically induced axonal injury by rapid posthypothermic rewarming and attenuation of axonal change by cyklosporin A. J Neurosurg 2001;94:493 /8. Gabriel A, Klussmann FW, Ingelmund P. Rapid temperature changes induce adenosine-mediated depression of synaptic transmission in hippocampal slices from rats (non-hibernators) but not in slices from golden hamsters (hibernators). Neuroscience 1998;86:67 /77. Glew MJ, Safar P, Stremple P. A simple survival model of volume-controlled hemorrhagic shock in awake rats. Resuscitation 1991;21:247 /57. Lilly MP, Gala GJ, Carlson DE, Sutherland BE, Gann DS. Saline resuscitation after fixed-volume hemorrhage. Role of resuscitation volume and rate of infusion. Ann Surg 1992;216:161 /71. Ward PA, Lentsch AB. The acute inflammatory response and its regulation. Arch Surg 1999;134:666 /9. Ayala A, Wang P, Ba ZF, Perrin MM, Ertel W, Chaudry IH. Differential alterations in plasma IL-6 and TNF levels after trauma and haemorrhage. Am J Physiol 1991;260:R167 /71. Jiang JX, Bahrami S, Leichtfried G, Redl H, Oehlinger W, Schlag G. Kinetics of endotoxin and tumor necrosis factor appearance in portal and systemic circulation following haemorrhagic shock in rats. Ann Surg 1995;221:100 /6. Kentner R, Rollwagen FM, Prueckner S, Behringer W, Wu S, Stezoski J, et al. Effects of mild hypothermia on survival and serum cytokines in uncontrolled hemorrhagic shock in rats. Shock 2002;17:521 /6. Volk T, Schmutzler M, Engelhardt L, Do¨cke W-D, Volk H-D, Konertz W, et al. Influence of aminosteroid and lucocorticoid treatment on inflammation and immune function during cardiopulmonary bypass. Crit Care Med 2001;29:2137 /42. Suter PM, Suter S, Girardin E, Roux-Lombard P, Grau GE, Dayer JM. High bronchoalveolar levels of tumor necrosis factor and its inhibitors, interleukin-1, interferon, and elastase, in patients with adult respiratory syndrome after trauma, shock, or sepsis. Am Rev Respir Dis 1992;145:1016 /22. Wu X, Prueckner S, Rollwagen F, Kentner R, Stezoski J, Kochanek PM, et al. Gut damage during haemorrhagic shock: effects on survival of oral or enteral interleukin-6. Shock 2001;16:449 /53. Davis JW, Mackersie RC, Holbrook TL, Hoyt DB. Base deficit as an indicator of significant abdominal injury. Ann Emer Med 1991;20:842 /4. Mizock BA, Falk JL. Lactic acidosis in critical illness. Crit Care Med 1992;20:80 /93. Johnston TD, Chen Y, Reed RL, II. Functional equivalence of hypothermia to specific clotting factor deficiencies. J Trauma 1994;37:413 /7. Ning XH, Xu CS, Song YC, Xiao Y, Hu YJ, Lupinetti FM. Hypothermia preserves function and signaling for mitochondrial biogenesis during subsequent ischaemia. Am J Physiol 1998;43:H786 /93.
P. Vaagenes et al. / Resuscitation 58 (2003) 103 /112
 Szabo´ C, Thiemermann C. Invited opinion: role of nitric oxide in haemorrhagic, traumatic, and anaphylactic shock and thermal injury. Shock 1994;2:145 /55.  Sakamoto KI, Fujisawa H, Koizumi H, Tsuchida E, Ito H, Sadamitsu D, et al. Effects of mild hypothermia on nitric oxide synthesis following contusion trauma in the rat. J Neurotrauma 1997;14:349 /53.  Gundersen Y, Corso CO, Leiderer R, Do¨rger M, Lilleaasen P, Aasen AO, et al. The nitric oxide donor sodium nitroprusside protects against hepatic microcirculatory dysfunction in early endotoxaemia. Intensive Care Med 1998;24:1257 /63.  Wang P, Zweier JL. Measurement of nitric oxide and peroxynitrite generation in the postischaemic heart. Evidence for perox-
ynitrite-mediated reperfusion injury. J Biol Chem 1996;271:29223 /30.  Kausalya S, Nath J. Interactive role of nitric oxide and superoxide anion in neutrophil-mediated endothelial cell injury. J Leukocyte Biol 1998;64:185 /91.  Beckman JS, Koppenol WH. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am J Physiol 1996;271:C1424 /37.  Zacharowski K, Olbrich A, Cuzzocrea S. Membrane-permeable radical scavenger, tempol, reduces multiple organ injury in a rodent model of Gram-positive shock. Crit Care Med 2000;28:1953 /61.