Secondary Logo

Journal Logo

Basic Science Aspects

Hyperoxia or Therapeutic Hypothermia During Resuscitation from Non-Lethal Hemorrhagic Shock in Swine

Nussbaum, Benedikt L.∗,†; Stenzel, Tatjana; Merz, Tamara; Scheuerle, Angelika; McCook, Oscar; Wachter, Ulrich; Vogt, Josef A.; Matallo, José; Gässler, Holger§; Gröger, Michael; Matejovic, Martin||; Calzia, Enrico; Lampl, Lorenz§; Georgieff, Michael; Möller, Peter; Asfar, Pierre¶,∗∗; Radermacher, Peter; Hafner, Sebastian∗,†

Author Information
doi: 10.1097/SHK.0000000000000884



Recently, in a model of prolonged, near lethal hemorrhagic shock, we reported that 22 h of hyperoxia ameliorated the inflammatory response and improved organ function, whereas mild hypothermia for 22 h increased barrier dysfunction and vascular leakage, thus, lacking therapeutic benefit (1). The salutary effects of hyperoxia are in good agreement with antecedent studies in models of hemorrhagic shock (2–4), whereas the detrimental properties of therapeutic hypothermia contrast with several studies reporting anti-inflammatory effects (5, 6) and improved survival (7–9). However, an increased and/or prolonged inflammatory response has also been reported for hypothermia (10, 11). Of note, various issues of therapeutic hypothermia comprising time point, duration, degree, and rate of cooling substantially vary among the available studies and are still under debate (12). Different aspects of timing most likely contributed to the variable results of our report, and more studies are required to define optimal targets for hyperoxia and hypothermia (13). Limitations of our previous study were the severity of the prolonged shock, the absence of rewarming after hypothermia, or lacking return to standard ventilation following hyperoxia. Therefore, the present follow-up study investigated the effects of shorter exposure (12 h) to hyperoxia or hypothermia during resuscitation from mitigated porcine hemorrhagic shock with subsequent return to standard ventilation or rewarming, respectively. We hypothesized that hyperoxia is still safe and effective in improving organ function, however, to a lesser extent due to a lower accumulated oxygen debt during less severe shock, and that shorter therapeutic hypothermia and rewarming would avoid harmful effects.



The study was approved by the University of Ulm Animal Care Committee and the Federal Authorities for Animal Research (animal experiment number 1087, date of approval by the Regierungspräsidium Tübingen: 08.05.2012). The experiments were performed in adherence with the National Institute of Health Guidelines on the Use of Laboratory Animals and the European Union “Directive 2010/63/EU on the protection of animals used for scientific purposes.” Thirty Bretoncelles-Meishan-Willebrand pigs without any cardiovascular comorbidities with a median age of 14.5 months (interquartile range (IQR): 13 to 16 months) and a median weight of 68 kg (IQR: 61–73 kg) were recruited for the study. As described previously (1), this pig strain is characterized by a reduced von Willebrand factor activity, and thus, more closely mimics the human coagulation system, as porcine blood typically is hypercoagulable compared with human blood. Two pigs died during or shortly after shock and one pig suffered from pneumonia prior to the experiment. These three animals were excluded from data analysis, and all subsequent data refer to the 27 pigs included in the final analysis.

Anesthesia, surgical instrumentation, measurements, and data analysis were identical to our previous publication (1). A detailed method description is provided in the supplemental digital content 1,

Experimental protocol

Figure 1 illustrates the experimental protocol. Briefly, hemorrhagic shock was induced by passive removal of 30% of the calculated blood volume. Mean arterial blood pressure (MAP) was titrated to 40 mm Hg for 3 h. At the end of shock, animals were randomized into three groups: the control, hyperoxia, and hypothermia group (each n = 9). While the control group received standard resuscitation for 23 h (see supplemental digital content 1,, providing a detailed description of methods), the hyperoxia group was ventilated with an inspiratory oxygen fraction (FIO2) of 100% within the first 12 h of resuscitation, and the hypothermia group was actively cooled to a core temperature of 34°C, respectively. After 12 h, standard ventilation ensued in the hyperoxia group and hypothermic pigs were actively rewarmed.

Fig. 1
Fig. 1:
Experimental protocol.



All animals survived during the 23 h of resuscitation in the hyperoxia group and only one animal died in the control group, whereas six animals prematurely died in the hypothermia group, resulting in a significantly higher mortality (Fig. 2).

Fig. 2
Fig. 2:
Survival curves.

Hemodynamics and acid–base status

The amounts of shed blood were comparable between groups (supplemental digital content 2—Table 1, Parameters of systemic and pulmonary hemodynamics, calorimetry, acid base balance, and glucose metabolism are demonstrated in Table 1. At the end of the hyperoxic or hypothermic treatment, hemodynamics were comparable between the hyperoxic and the control group, whereas MAP was 38% and cardiac output 49% lower (P = 0.008 and P = 0.025, respectively) in the hypothermic animals compared with controls despite higher overall norepinephrine consumption (P = 0.035, supplemental digital content 2—Table 1, Hemorrhagic shock increased lactate and decreased base excess (BE) in all groups. Consistent with the impaired circulation, lactate was significantly higher (11.0 (6.6;13.6) mmol L−1 vs. 1.0 (0.8;1.5) mmol L−1) and BE significantly lower (−15.4 (−22.3; −7.5) mmol L−1 vs. 3.8 (0.3;5.2) mmol L−1) after 12 h of resuscitation in the hypothermic group compared to control animals. At 23 h of resuscitation and after rewarming, MAP and cardiac output only slightly increased, BE remained decreased, and lactate levels remained elevated in the hypothermic group. In contrast, systemic hemodynamics and acid–base balance returned to baseline values in the hyperoxic group. Hemoglobin levels increased in all groups compared with baseline but were 55% higher (P = 0.005) in the hypothermic animals compared to control pigs after 12 h.

Table 1
Table 1:
Parameters of systemic and pulmonary hemodynamics, calorimetry, blood gases, acid–base balance, and glucose metabolism.

Visceral organ function

Table 2 illustrates parameters of organ function and injury. Cardiac troponin I strongly increased in all groups after hemorrhagic shock without significant inter-group differences. Horowitz index was significantly higher in the hyperoxia group after 12 h of resuscitation. Otherwise, hyperoxia did neither improve nor worsen organ function compared with the controls. In line with the compromised systemic hemodynamics in the hypothermia group, right kidney artery flow was significantly lower after 12 and 23 h post-shock. Consistently, urine output was reduced (P = 0.106) and creatinine clearance was 66% lower than in the control group after 12 h (P = 0.008). Hemorrhagic shock resulted in renal venous lactic acidosis in all treatment groups. While acid–base disturbances resolved during resuscitation in the hyperoxic and control group, renal venous lactic acidosis persisted in the hypothermic animals until 23 h of resuscitation. Increased plasma NGAL levels after hemorrhagic shock did not show statistically significant inter-group differences.

Table 2
Table 2:
Parameters of visceral organ function and injury


Table 3 shows increased markers of systemic inflammation and oxidative/nitrosative stress in all three groups after hemorrhagic shock. Plasma levels of IL-6 were higher (P = 0.067) at the end of the hypothermic treatment compared to the control group. In the hyperoxic animals, inflammatory parameters were comparable to controls, whereas nitrite/nitrate levels were 56% lower (P = 0.043) at the end of the experiment.

Table 3
Table 3:
Parameters of inflammation and oxidative/nitrosative stress

Supplemental digital content 3 (Figure 1, and supplemental digital content 4 (Figure 2, demonstrate the results of the western blots and EMSAs of renal tissue. Expression of IκBα, Bcl-xL, and activation of NF-κB did not significantly differ between groups. HO-1 expression was lower in the hyperoxia and hypothermia groups, however, overall expression levels in all three groups did not strongly differ from native animals. Renal iNOS expression was higher in the hyperoxia-treated pigs (P = 0.005).

Immunohistochemical analysis of kidneys did not reveal any significant differences between the hyperoxia and the control group (Fig. 3). In contrast, both extravascular albumin accumulation and 3-nitrotyrosine formation were significantly increased, whereas cystathionine γ-lyase (CSE) expression was lowest (P = 0.023) in the hypothermia group compared to controls.

Fig. 3
Fig. 3:
Immunohistochemistry of kidney specimens.


Supplemental digital content 5 (Table 2, summarizes the hemostaseological results. ROTEM analysis did not reveal major alterations of coagulation after hemorrhage and resuscitation in all groups. vWF activity significantly increased in all groups and was highest after 12 h of resuscitation in the hypothermia group (P = 0.032). Platelets decreased after 23 h of resuscitation without any significant inter-group differences. Thrombin–antithrombin complexes did not show any significant inter-group differences either.


We previously reported beneficial effects of 22 h of hyperoxic treatment but detrimental properties of therapeutic hypothermia during resuscitation from near-lethal hemorrhagic shock (1). Limitations of the study comprised the shock severity and the lack of rewarming or return to standard ventilation, respectively. Thus, the present follow-up study investigated the effects of shorter exposure (12 h) to hyperoxia or mild therapeutic hypothermia with subsequent return to standard ventilation or rewarming following less severe hemorrhagic shock. The main findings were that hyperoxia proved to be equally safe but did not further improve organ dysfunction. Conversely, (shorter) therapeutic hypothermia with subsequent rewarming exerted deleterious effects resulting in higher mortality in line with our previous results (1).

Effects of hyperoxia

Several aspects need to be addressed in order to explain the lacking beneficial properties of hyperoxia in the present study compared to our antecedent report (1): The exposure to hyperoxia was shorter (12 h vs. 22 h) in the current treatment protocol. However, kidney function already improved after 12 h of resuscitation in our previous study (1), and much shorter intervals of hyperoxia have yielded positive results both in rodents (4, 14) and porcine models (2). More importantly, the hemorrhagic shock was attenuated in comparison to our previous report (MAP of 40 mm Hg for 3 h vs. 35 mm Hg for 4 h). The deteriorated circulation during shock leads to an accumulating oxygen debt, and the degree of oxygen debt determines survival (15). Rapid repayment of oxygen debt represents an important feature of shock resuscitation strategies (16). We did not specifically quantify oxygen debt in the present study, but it has been demonstrated that metabolic parameters—in particular base deficit—strongly correlate with the amount of oxygen debt (17). BE was significantly lower at the end of shock in our previous study (both across all study animals −5.2 (−7.8; −1.8) vs. −1.6 (−4.4; 0.3) and among the hyperoxic groups only −5.4 (−9.7; −2.9) vs. −0.6 (−5.2; 2.1)), thus, strongly indicating an increased oxygen debt. Thus, it is tempting to speculate that hyperoxia represents a more effective therapeutic instrument with increasing shock severity by accelerating oxygen debt repayment. During resuscitation from less severe hemorrhagic shock, improvement of oxygen delivery by standard intensive care measures, i.e., restoration of the circulating blood volume and consequently cardiac output with adequate fluid administration might be sufficient to repay oxygen debt, and therefore, hyperoxia might not harbor further advantages. Consistently, hyperoxic treatment improved survival in awake rats in hemorrhagic shock without fluid administration (18, 19), whereas hyperoxia did not substantially affect survival in hemorrhagic shock models including early fluid resuscitation regimens (7, 14, 20).

Despite the lacking beneficial properties of hyperoxia in the present study, hyperoxia represented a safe approach and did not exert any harmful effects such as adverse oxygen-induced pulmonary reactions termed hyperoxic acute lung injury (21, 22). In our model, neither lung mechanics nor gas exchange deteriorated, which is in line with findings after even longer exposure to hyperoxia (22–24 h) during hemorrhagic (1) or septic shock (23). Of note, the Horowitz index was highest in the hyperoxia group after 12 h of resuscitation.

Plasma levels of inflammatory cytokines and 8-isoprostane as well as renal nitrotyrosine formation did not significantly differ between the hyperoxia and the control group. Interestingly, nitrite/nitrate levels were significantly lower in the hyperoxic animals after 23 h. A possible NO quenching effect by superoxide anions in the hyperoxia group resulting in the generation of peroxynitrite as cause for the reduced nitrite/nitrate levels seems rather unlikely, since nitrotyrosine formation did not significantly differ between the hyperoxia and the control group. Moreover, in line with our previous study (1), renal iNOS expression was significantly higher in the hyperoxic group. Effects of hyperoxia on inflammation and oxidative/nitrosative stress are controversial depending on the duration of hyperoxia and on the shock model: hyperoxia reduced the inflammatory cytokines in septic or hemorrhagic shock (1, 24) in swine. In contrast, parameters of inflammation or oxidative/nitrosative stress in rodents were augmented in renal or cerebral tissue following hyperoxic treatment (3, 4).

Effects of hypothermia

Mild therapeutic hypothermia significantly increased mortality. Systemic hemodynamics were substantially compromised necessitating high doses of norepinephrine. Consequently, BE was lowest, and lactate levels highest. Kidney perfusion was significantly decreased resulting in the lowest urine output and creatinine clearance. Rewarming after 12 h only partially mitigated the pathophysiologic derangements and could not restore baseline values. In line with our previous report (1), significantly augmented hemoglobin at the end of the hypothermic phase and increased renal extravascular albumin indicate higher vascular endothelial permeability with fluid shifts into the extravascular space as the main cause for the circulatory impairment and subsequent organ dysfunction. Hypothermia-induced extravasation of fluids has previously been demonstrated (25, 26). However, following hemorrhagic shock, both increased and decreased vascular permeability have been reported (27, 28).

Renal CSE expression—important for the preservation of kidney function during shock (29)—was significantly lower in the hypothermia group. Both reduced expression of CSE as well as decreased vascular integrity and the amount of extravascular albumin have been demonstrated to be associated with kidney dysfunction (30). In contrast, pretreatment hypothermia mitigated organ dysfunction, attenuated the hemorrhagic shock-induced decrease in renal CSE expression, and was associated with lower plasma creatinine values (10). Moreover, hyperoxia-induced increased CSE expression and reduced albumin extravasation coincided with lower plasma creatinine levels during resuscitation from hemorrhagic shock (1). Interestingly, renal CSE levels and extravascular albumin did not significantly differ between the hypothermic and the control group in our antecedent study, suggesting that rewarming may even have contributed to aggravated vascular leakage and organ dysfunction.

With respect to the inflammatory response during hemorrhage and resuscitation, both anti-inflammatory effects (5, 6) as well as increased inflammatory parameters have been reported (10, 11) after application of hypothermia. In the present study, plasma levels of IL-6 were higher after 12 h of resuscitation in the hypothermic group, while other inflammatory parameters remained largely unaffected. In contrast to our previous report (1), nitrotyrosine formation was significantly higher in the hypothermic animals, indicating increased oxidative and nitrosative stress during rewarming. Consistently, rewarming of rats after severe hypothermia revealed significantly decreased antioxidant parameters (31).

The detrimental effects of mild therapeutic hypothermia and subsequent rewarming following hemorrhagic shock are consistent with our previous findings during more severe, near-lethal shock (1). In contrast, various other studies described beneficial effects of induced hypothermia (6, 8, 9) or at least safe administration (32). In order to reconcile the divergent results, “issues of timing” seem to play a crucial rule (13). Studies vary substantially with respect to the time point of initiation, duration, rate, degree, and technique of cooling (12). Positive effects have mainly been demonstrated with shorter intervals of hypothermia (6, 9). Interestingly, 12 h of hypothermia—as used in the present study—significantly improved survival after hemorrhage, when cooling was already initiated during shock, and target temperature reached faster (8). In contrast, start of hypothermia for 12 h after shock and resuscitation with a similar rate of cooling prolonged the inflammatory response (11). Moreover, differences of the shock models regarding the duration and degree of hemorrhage (and associated trauma) most likely further contribute to the discrepant results.


The lacking beneficial effects of hyperoxia following hemorrhagic shock in the present study together with the previously described organ-protective effects after near-lethal shock suggest a dependence of the effectiveness of hyperoxia from shock severity. Importantly, hyperoxia did not exert any adverse effects and proved to be safe in both studies.

In line with our antecedent report, mild therapeutic hypothermia was confirmed to be detrimental. Different aspects of “timing and dosing” of hypothermia most likely account for the discrepant findings compared with previously described beneficial effects.


The authors thank Andrea Seifritz, Anja Gröger, Bettina Stahl, Rosemarie Mayer, Marina Fink, Tanja Schulz, Rosa Engelhardt, and Ingrid Eble for their skillful technical assistance.


1. Knöller E, Stenzel T, Broeskamp F, Hornung R, Scheuerle A, McCook O, Wachter U, Vogt JA, Matallo J, Wepler M, et al. Effects of hyperoxia and mild therapeutic hypothermia during resuscitation from porcine hemorrhagic shock. Crit Care Med 2016; 44 5:e264–e277.
2. Meier J, Kemming GI, Kisch-Wedel H, Blum J, Pape A, Habler OP. Hyperoxic ventilation reduces six-hour mortality after partial fluid resuscitation from hemorrhagic shock. Shock 2004; 22 3:240–247.
3. Efrati S, Berman S, Ben Aharon G, Siman-Tov Y, Averbukh Z, Weissgarten J. Application of normobaric hyperoxia therapy for amelioration of haemorrhagic shock-induced acute renal failure. Nephrol Dial Transplant 2008; 23 7:2213–2222.
4. Blasiole B, Bayr H, Vagni VA, Janesko-Feldman K, Cheikhi A, Wisniewski SR, Long JB, Atkins J, Kagan V, Kochanek PM. Effect of hyperoxia on resuscitation of experimental combined traumatic brain injury and hemorrhagic shock in mice. Anesthesiology 2013; 118 3:649–663.
5. Coyan GN, Moncure M, Thomas JH, Wood JG. Induced hypothermia during resuscitation from hemorrhagic shock attenuates microvascular inflammation in the rat mesenteric microcirculation. Shock 2014; 42 6:518–524.
6. Fröhlich M, Hildebrand F, Weuster M, Mommsen P, Mohr J, Witte I, Raeven P, Ruchholtz S, Flohé S, van Griensven M, et al. Induced hypothermia reduces the hepatic inflammatory response in a swine multiple trauma model. J Trauma Acute Care Surg 2014; 76 6:1425–1432.
7. Takasu A, Stezoski SW, Stezoski J, Safar P, Tisherman SA. Mild or moderate hypothermia, but not increased oxygen breathing, increases long-term survival after uncontrolled hemorrhagic shock in rats. Crit Care Med 2000; 28 7:2465–2474.
8. Wu X, Kochanek PM, Cochran K, Nozari A, Henchir J, Stezoski SW, Wagner R, Wisniewski S, Tisherman SA. Mild hypothermia improves survival after prolonged, traumatic hemorrhagic shock in pigs. J Trauma 2005; 59 2:291–301.
9. George ME, Mulier KE, Beilman GJ. Hypothermia is associated with improved outcomes in a porcine model of hemorrhagic shock. J Trauma 2010; 68 3:662–668.
10. Gröger M, Scheuerle A, Wagner F, Simon F, Matallo J, McCook O, Seifritz A, Stahl B, Wachter U, Vogt JA, et al. Effects of pretreatment hypothermia during resuscitated porcine hemorrhagic shock. Crit Care Med 2013; 41 7:e105–e117.
11. Horst K, Eschbach D, Pfeifer R, Relja B, Sassen M, Steinfeldt T, Wulf H, Vogt N, Frink M, Ruchholtz S, et al. Long-term effects of induced hypothermia on local and systemic inflammation—results from a porcine long-term trauma model. PLoS ONE 2016; 11 5:e0154788.
12. Hildebrand F, Radermacher P, Ruchholtz S, Huber-Lang M, Seekamp A, Flohé S, van Griensven M, Andruszkow H, Pape H-C. Relevance of induced and accidental hypothermia after trauma-haemorrhage—what do we know from experimental models in pigs? Intensive Care Med Exp 2014; 2 1:16.
13. Tisherman SA. Targeted oxygen and temperature management during hemorrhagic shock: is it all in the timing? Crit Care Med 2016; 44 5:1024–1025.
14. Takasu A, Iwamoto S, Ando S, Minagawa Y, Kashiba M, Yamamoto Y, Sakamoto T. Effects of various concentrations of inhaled oxygen on tissue dysoxia, oxidative stress, and survival in a rat hemorrhagic shock model. Resuscitation 2009; 80 7:826–831.
15. Rixen D, Siegel JH. Bench-to-bedside review: oxygen debt and its metabolic correlates as quantifiers of the severity of hemorrhagic and post-traumatic shock. Crit Care 2005; 9 5:441–453.
16. Barbee RW, Reynolds PS, Ward KR. Assessing shock resuscitation strategies by oxygen debt repayment. Shock 2010; 33 2:113–122.
17. Rixen D, Raum M, Holzgraefe B, Sauerland S, Nagelschmidt M, Neugebauer EA. Shock and Trauma Study Group: a pig hemorrhagic shock model: oxygen debt and metabolic acidemia as indicators of severity. Shock 2001; 16 3:239–244.
18. Adir Y, Bitterman N, Katz E, Melamed Y, Bitterman H. Salutary consequences of oxygen therapy on the long-term outcome of hemorrhagic shock in awake, unrestrained rats. Undersea Hyperb Med 1995; 22 1:23–30.
19. Crippen D, Safar P, Porter L, Zona J. Improved survival of hemorrhagic shock with oxygen and hypothermia in rats. Resuscitation 1991; 21 (2–3):271–281.
20. Takasu A, Prueckner S, Tisherman SA, Stezoski SW, Stezoski J, Safar P. Effects of increased oxygen breathing in a volume controlled hemorrhagic shock outcome model in rats. Resuscitation 2000; 45 3:209–220.
21. Hafner S, Beloncle F, Koch A, Radermacher P, Asfar P. Hyperoxia in intensive care, emergency, and peri-operative medicine: Dr. Jekyll or Mr. Hyde? A 2015 update. Ann Intensive Care 2015; 5 1:42.
22. Kallet RH, Branson RD. Should oxygen therapy be tightly regulated to minimize hyperoxia in critically ill patients? Respir Care 2016; 61 6:801–817.
23. Barth E, Bassi G, Maybauer DM, Simon F, Gröger M, Öter S, Speit G, Nguyen CD, Hasel C, Möller P, et al. Effects of ventilation with 100% oxygen during early hyperdynamic porcine fecal peritonitis. Crit Care Med 2008; 36 2:495–503.
24. Hauser B, Barth E, Bassi G, Simon F, Gröger M, Öter S, Speit G, Ploner F, Möller P, Wachter U, et al. Hemodynamic, metabolic, and organ function effects of pure oxygen ventilation during established fecal peritonitis-induced septic shock. Crit Care Med 2009; 37 8:2465–2469.
25. Farstad M, Heltne JK, Rynning SE, Lund T, Mongstad A, Eliassen F, Husby P. Fluid extravasation during cardiopulmonary bypass in piglets—effects of hypothermia and different cooling protocols. Acta Anaesthesiol Scand 2003; 47 4:397–406.
26. Hammersborg SM, Brekke HK, Haugen O, Farstad M, Husby P. Surface cooling versus core cooling: comparative studies of microvascular fluid- and protein-shifts in a porcine model. Resuscitation 2008; 79 2:292–300.
27. Norio H, Takasu A, Kawakami M, Saitoh D, Sakamoto T, Okada Y. Rapid body cooling by cold fluid infusion prolongs survival time during uncontrolled hemorrhagic shock in pigs. J Trauma 2002; 52 6:1056–1061.
28. Childs EW, Udobi KF, Hunter FA. Hypothermia reduces microvascular permeability and reactive oxygen species expression after hemorrhagic shock. J Trauma 2005; 58 2:271–277.
29. McCook O, Radermacher P, Volani C, Asfar P, Ignatius A, Kemmler J, Möller P, Szabó C, Whiteman M, Wood ME, et al. H2S during circulatory shock: some unresolved questions. Nitric Oxide 2014; 41:48–61.
30. Stenzel T, Weidgang C, Wagner K, Wagner F, Gröger M, Weber S, Stahl B, Wachter U, Vogt J, Calzia E, et al. Association of kidney tissue barrier disrupture and renal dysfunction in resuscitated murine septic shock. Shock 2016; 46:398–404.
31. Alva N, Palomeque J, Carbonell T. Oxidative stress and antioxidant activity in hypothermia and rewarming: can RONS modulate the beneficial effects of therapeutic hypothermia? Oxid Med Cell Longev 2013; 2013:957054.
32. Weuster M, Mommsen P, Pfeifer R, Mohr J, Ruchholtz S, Flohé S, Fröhlich M, Keibl C, Seekamp A, van Griensven M, et al. Induced hypothermia does not harm hemodynamics after polytrauma: a porcine model. Mediators Inflamm 2015; 2015 4:829195–829211.

3-nitrotyrosine; barrier dysfunction; extravascular albumin; heme oxygenase-1; inducible nitric oxide synthase; oxidative stress

Supplemental Digital Content

Copyright © 2017 by the Shock Society