Despite substantial progress in intensive care medicine in the last years, systemic inflammatory response syndrome (SIRS) still remains a significant clinical problem resulting from infective or non-infective inflammatory stimuli, and leading to an excessive liberation of cytotoxic mediators responsible for cellular and organ dysfunction . Typically, endothelial damage occurs and results in a capillary leakage syndrome followed by tissue oedema . In addition, pro-inflammatory cytokines, such as tumour necrosis factor alpha (TNF-α) and interleukin-1 (IL-1), produce large amounts of nitric oxide (NO) by activating the inducible form of nitric oxide synthetase (iNOS) . Nitric oxide then activates soluble guanylate cyclase, which increases intracellular cyclic guanosine 3′,4′-monophosphate and leads to significant relaxation of vascular smooth muscle cells . In sepsis, however, vascular adenosine 5′-triphosphate (ATP)-regulated K+ channels are activated and lead to a reduced entry of Ca2+ through voltage-gated Ca2+ channels. In addition, activation of KATP channels leads to relaxation of resistance vessels and causes cardiac myocytes to lose their ability to contract sufficiently, thus playing an important role in the pathogenesis of cardiovascular failure in sepsis .
In contrast to other shock forms, septic shock is characterized by two distinct haemodynamic phases: the early hyperdynamic stage is characterized by a fall in systemic vascular resistance index (SVRI) followed by a reduction in mean arterial pressure (MAP). Via sympathetic reflex mechanisms, heart rate (HR) and cardiac output (CO) are elevated, and to a certain extent, compensate for the decrease in arterial pressure and mostly increased tissue oxygen demands. When sepsis progresses a transition to the moribund hypodynamic phase may occur. These derangements in the systemic circulation are often complicated by pulmonary hypertension and typically associated with increased mortality rates . Despite the clinical significance of studying the pathophysiological mechanisms responsible for circulatory failure under these circumstances, there are only a few large animal models available, reliably reproducing and maintaining this late stage of septic shock. Most studies with the valid model of ovine endotoxaemia and sepsis, respectively, were based on a hyperdynamic circulation, and focused on therapeutic approaches to reverse this condition [7-9]. Since these investigations were performed with low-dose ETX (10 ng kg−1 min−1), our hypothesis was that higher dosages could be useful to induce a hypodynamic circulation, as it is known from septic patients. Using an ovine model, we studied the cardiopulmonary response following a continuous ETX infusion at incrementing rates.
After approval by the local Government Animal Research Committee, eight adult sheep with a mean weight of 40 ± 2 kg were instrumented for chronic study. Anaesthesia was induced with intramuscular xylazine 2% (Xylazin®, 0.15 mg kg−1; CEVA Tiergesundheit GmbH, Düsseldorf, Germany) and ketamine (Ketanest 50®, 15 mg kg−1; Parke-Davis, Freiburg, Germany). Then, the ewes were chronically instrumented with an indwelling pulmonary artery catheter positioned percutaneously via an introducer sheath into the right jugular vein (8.5 Fr. Catheter Introducer Set; pvb Medizintechnik GmbH, Kirchseeon, Germany; and 7.5 Fr. Edwards Swan Ganz; Edwards Critical Care Division, Irvine, CA, USA). For continuous arterial pressure measurements, a femoral arterial catheter (18-G Leader Cath®; Vygon Ltd., Aachen, Germany) was placed into the left femoral artery. During the instrumentation, anaesthesia was maintained with a continuous intravenous (i.v.) infusion of propofol (Disoprivan®, 4-5 mg kg−1h−1; AstraZeneca, Schwetzigen, Germany). After the instrumentation, the ewes received a single dose infusion of ceftriaxone (Rocephin®; Hoffmann-La Roche AG, Grenzach-Wyhlen, Germany). To provide sufficient fluid challenge, a continuous i.v. infusion of Ringer's lactate solution was adjusted to keep central venous pressure (CVP) and PCWP at baseline values ±3 mmHg (2-6 mL kg−1h−1). After 24h for recovery, catheters were connected to pressure transducers (DTX® pressure transducer; Ohmeda Ltd. and Co. KG, Erlangen, Germany) and a physiological recorder (Hellige Servomed®; Hellige Ltd., Freiburg, Germany), to monitor haemodynamic variables of the systemic and pulmonary circulation. Cardiac output measurements were performed according to the thermodilution technique, using triplicate 10 mL injections of 2-5°C cold saline solution (9520 A cardiac output computer; Edward Lifescience, Irvine, USA). In addition, core body temperature was determined with the pulmonary artery catheter. SVRI and pulmonary vascular resistance index (PVRI) were calculated using standard equations. Arterial lactate concentration, mixed venous and arterial oxygen content, haemoglobin, and haemoglobin oxygen saturation were analysed at the measured body temperature with an ABL 625® (Radiometer Copenhagen; Copenhagen, Denmark). Oxygen delivery (DO2) was calculated from the product of CO and the systemic arterial oxygen content, and oxygen consumption (VO2) as the product of CO and arteriovenous oxygen content difference. Oxygen extraction rate (O2-ER) was derived from arteriovenous oxygen content difference/arterial oxygen content. All measurements were performed in awake sheep that were spontaneously breathing room air. Throughout the experiments, the animals were housed and studied in metabolic cages with free access to water and food.
All animals were included in the study. Their baseline HR was <100 beats min−1, body temperature <40°C and arterial lactate concentration <1 mmol L−3. Following this baseline measurement, healthy sheep were subjected to a continuous ETX infusion (Salmonella typhosa; Sigma Chemicals, Deisenhofen, Germany) started with 10 ng kg−1 min−1 and doubled every hour for seven times. Haemodynamic and oxygen transport variables and arterial lactate concentrations were obtained at baseline and every hour during the study. At the end of the experiment, the surviving ewes were anaesthetized with a bolus infusion of propofol (4 mg kg−1; Disoprivan®) and killed with an overdose of potassium chloride.
Data are expressed as mean ± SEM. For statistical analysis, Sigma Stat® 2.03 software (SPSS Inc., Chicago, IL, USA) was used. After testing for normal distribution (Kolmogorov-Smirinov), a one-way analysis of variance (ANOVA) for repeated measurements with appropriate post hoc comparisons (Student-Newman-Keuls) was performed. P < 0.05 was considered significant.
Effects of ETX infusion on cardiopulmonary performance
Following 4 h of endotoxaemia, two sheep died due to circulatory failure and were excluded from data analysis. In the ewes surviving the entire experiment (n = 6), ETX infusion resulted in a progressive decrease in MAP (Fig. 1a) and CO (Fig. 2a). This was accompanied by a significant increase in HR (Fig. 2b) and an initially augmented SVRI (Fig. 1b). Following a dose of 80 ng kg−1 min−1, the maximum decrease in CO occurred (Fig. 2a), and was associated with the greatest increase in SVRI (Fig. 1b). However, doses > 80 ng kg−1 min−1 caused a successive decrease in SVRI, accompanied by severe systemic hypotension, tachycardia and low CO. After 1 h of ETX challenge both mean pulmonary arterial pressure (MPAP) (Fig. 3a) and PVRI (Fig. 3b) were augmented, and remained significantly increased throughout the experiment.
Effects of ETX infusion on global oxygen transport
In a dose-dependent manner, ETX infusion resulted in a decrease in both DO2 and VO2. This was associated with an early increase in O2-ER. Following ETX doses >40 ng kg−1 min−1, O2-ER successively decreased (Table 1). In addition, ETX infusion caused an increase in arterial lactate concentration (Fig. 4a) and core body temperature (Fig. 4b).
The major finding of this present study is that with incrementing ETX doses, a hypodynamic circulation, accompanied by pulmonary hypertension and tissue dysoxia, could reliably be produced and maintained in sheep.
In a dose-dependent manner, S. typhosa ETX decreased both MAP and CO. Since fluid resuscitation is a hallmark in haemodynamic support of septic patients, we adjusted the infusion rate to keep CVP and PCWP at baseline values to guarantee adequate fluid challenge. However, the reflectory increase in HR did not result in an elevated CO, suggesting that ETX infusion decreased ventricular stroke volume, most likely due to myocardial depression. Notably, pilot studies investigating haemodynamic effects of a more aggressive fluid therapy resulted in an increase in CVP without restoring MAP and systemic blood flow. Therefore, we chose an approach maintaining cardiac filling pressures within physiologic ranges.
In the clinical setting, where the septic patient becomes refractory to fluid challenge, vasoactive drugs are usually administered, aiming at the preservation of sufficient organ perfusion and tissue oxygenation. Since preliminary data of our laboratory demonstrate that, adrenomedullin is useful to reverse the hypodynamic circulatory state to a more stable hyperdynamic circulation, thereby reducing mortality (data not presented), we feel that this model is useful to study new therapeutic approaches.
Similarly to the haemodynamic changes found in clinical practice, SVRI initially increased to compensate for the fall in CO . With ETX doses >80 ng kg−1 h−1 SVRI finally decreased and resulted in a detrimental fall in MAP. A reduction in both systemic blood flow and perfusion pressure in a state of increased oxygen requirements, such as SIRS, denotes an increased risk for organ hypoperfusion and cellular hypoxia. This was also observed in the presented study. The decrease in tissue oxygen uptake obviously resulted from the fall in systemic DO2. Most likely, systemic DO2 was fallen beyond the critical point of DO2, where systemic VO2 becomes dependent on DO2 and thus decreases if DO2 continues to fall. Such a mismatch between tissue oxygen supply on the one hand and oxygen demand on the other hand-which is often referred to as 'oxygen supply dependency' - typically results in an anaerobic metabolism and is followed by localized or generalized lactate production . This study demonstrated a significant increase in arterial lactate concentration following ETX doses >40 ng kg−1 min−1, confirming the occurrence of tissue hypoxia. Additionally, this was accompanied by an increase in core body temperature, most likely as a result of the systemic inflammatory response. Moreover, ETX challenge was associated with pulmonary hypertension, as seen in septic patients . Since septic shock still has a high mortality rate, its treatment is a continuing challenge for the intensive care physician. During the course of our study, two sheep died due to circulatory failure. The observed fall in CO was accompanied by a progressive decrease in SVRI, resulting in a substantial fall in systemic blood pressure. These haemodynamic changes were associated with a mismatch between DO2 and O2 demand and linked with systemic hypoxaemia. Interestingly, the sheep surviving the experiment did not undergo hypoxaemia, suggesting that tissue oxygenation plays a pivotal role in the preservation of cardiovascular functions in this model.
To date there are some reports on hypodynamic large animal models available, including the lethal sheep model described by Masouye and colleagues . Administering Escherichia coli endotoxin as a continuous infusion (40 ng kg−1 min−1) induced a hypotensive-hypodynamic circulation. However, this approach resulted also in systemic hypoxaemia and led to death within few hours. Although the efficacy of xanthine derivatives to improve outcome in endotoxin shock has been documented in other animal models , albifyllin failed to improve survival in this model . However, no study has ever been published, demonstrating that any therapeutic approach led to either a sustained improvement in cardiovascular functions or in survival. The 100% mortality rate in their model suggests that sheep are obviously more sensitive to endotoxin derived from E. coli than from S. typhosa. The advantage of our new approach is that this model is sensitive to inotropic agents and has only a mortality rate of 25%.
To our knowledge, this is the first study demonstrating that infusion of S. typhosa ETX at doses ≥80 ng kg−1 h−1 is useful to induce and maintain a hypodynamic circulation. Whether this condition resulted from cardiac failure due to excessive NO liberation, or whether it was directly linked to the high ETX concentrations needs to be investigated in future studies. Since severe haemodynamic alterations, as observed in our study, may lead to multiple organ failure  or even death , the underlying pathophysiology seems worthy of further detailed study. In this regard, it is important to establish a model, allowing to the study of key parameters being (a) responsible for cardiovascular failure in sepsis and SIRS, and (b) suitable to evaluate the efficacy of therapeutic approaches. The latter would include fluid resuscitation, blood-gas analysis and adequate monitoring of the cardiopulmonary performance. Since a pulmonary artery catheter is necessary to study these parameters in greater detail, it is almost impossible to do this accurately in small animal models. Nevertheless, the hypodynamic circulation is most commonly studied in rats [15,16] and rabbits . In this regard, it is also noteworthy, that the rat requires ETX doses higher than 4 mg kg−1 to achieve a cardiopulmonary response. This is 4 million times the amount of ETX needed for a haemodynamic response in healthy human volunteers . Similarly, it takes logarithmic differences in bacteria and ETX dosages to produce a haemodynamic response in pigs  and dogs [20,21]. Therefore, special care must be taken when transferring data obtained from such animal models to human beings. In this context, it appears to be appropriate to use animal models large enough to allow for monitoring and treatment strategies (including resuscitation). Further, the animal model should closely mimic human beings in the response and sensitivity to bacteria and endotoxin, respectively. When discussing the usefulness of primate sepsis models, it is noteworthy that the Rhesus monkey is almost completely refractory to ETX challenge, and therefore unacceptable to study endotoxin shock . Since the haemodynamic response to ETX and Pseudomas aeruginosa in man and sheep is nearly the same, and since the lungs of these animals are similar to human beings, the sheep has been used in the study of sepsis for more than three decades . However, previous studies with the ovine model primarily focused on the pathophysiology of hyperdynamic circulation [7-9].
The ovine model we have described here may help to study new treatment strategies for haemodynamic support in the common setting of a hypodynamic circulation, which may be applied to human beings in a clinical situation.
A limitation of our study is that we investigated effects during ovine endotoxaemia, and not during human sepsis. However, the effects on haemodynamic and oxygen transport variables following our approach are similar compared with the changes observed in the late stage of septic shock in human beings . In addition, we calculated DO2 and VO2 according to the Fick principle. Since both parameters are mathematically coupled, the possibility also exists that the reduction in VO2 could have been resulted from the decrease in CO. Alternatively, it would have been possible to determine VO2 independently from DO2, e.g. using calorimetry. However, since this technique requires sedation and intubation, which in turn alter haemodynamic variables, we did not use such measurements.
In summary, our results demonstrate that with ETX doses ≥80 ng kg−1 min−1, the clinical scenario of a hypodynamic circulation can easily and reliably be produced in sheep. This offers new options to study both the pathophysiology and the therapeutic approaches in a model closely mimicking the clinical situation known from patients with progressive sepsis and SIRS, respectively.
This study was supported by the Innovative Medizinische Forschung (IMF), University of Münster, Münster, Germany.
1. Latifi SQ, O'Riordan MA, Levine AD. Interleukin-10 controls the onset of irreversible septic shock. Infect Immun
2. Vallet B. Endothelial cell dysfunction and abnormal tissue perfusion. Crit Care Med
3. Suzuki N, Sakamoto A, Ogawa R. Effect of L-canavanine, an inhibitor of inducible nitric oxide synthase, on myocardial dysfunction during septic shock. J Nippon Med Sch
4. Mosi R, Seguin B, Cameron B, Amankwa L, Darkes MC, Fricker SP. Mechanistic studies on AMD6221: a ruthenium-based nitric oxide scavanger. Biochem Biophys Res Commun
5. Landry DW, Oliver JA. The ATP-sensitive K+
channel mediates hypotension in endotoxemia and hypoxic lactic acidosis in dog. J Clin Invest
6. Schoemaker WC, Appel PL, Kram HB, Bishop MH, Abraham E. Sequence of physiologic patterns in surgical septic shock. Crit Care Med
7. Meyer J, Booke M, Waurick R, Prien T, Van Aken H. Nitric oxide synthase inhibition restores vasopressor effects of norepinephrine in ovine hyperdynamic sepsis. Anesth Analg
8. Westphal M, Stubbe H, Bone HG, et al.
Hemodynamic effects of exogenous adrenomedullin in healthy and endotoxemic sheep. Biochem Biophys Res Commun
9. Scharte M, Meyer J, Van Aken H, Bone HG. Hemodynamic effects of terlipressin (a synthetic analog of vasopressin) in healthy and endotoxemic sheep. Crit Care Med
10. Sielenkämper AW, Yu P, Eichelbrönner O. Diaspirin crosslinked Hb and norepinephrine prevent the sepsis-induced increase in critical O2
delivery. Am J Physiol Heart Circ Physiol
11. Task Force of the American College of Critical Care Medicine, Society of Critical Care Medicine. Practice parameters for hemodynamic support of sepsis in adult patients in sepsis. Crit Care Med
12. Masouye K, Gunning K, Pittlet JF, Suter PM, Morel DR. Xanthine derivative HWA 138 attenuates hemodynamic and oxygen uptake dysfunction secondary to severe endotoxin shock in sheep. Circ Shock
13. Bahrami S, Yao YM, Shiga H, Leichtfried G, Redl H, Schlag G. Comparison of the efficacy of pentoxyfylline and albifyllin (HWA 138) on endotoxin-induced cytokine production, coagulation disturbances, and mortality. Shock
14. Landry DW, Oliver JA. The pathogenesis of vasodilatory shock. New Engl J Med
15. Koo DJ, Zhou M, Chaudry IH, Wang P. The role of adrenomedullin in producing differential hemodynamic responses during sepsis. J Surg Res
16. Ornan DA, Chaudry IH, Wang P. Pulmonary clearance of adrenomedullin is reduced during the late stage of sepsis. Biochim Biophys Acta
17. Ricard-Hibon A, Losser MR, Kong R, Beloucif S, Teisseire B, Payen D. Systemic pressure-flow reactivity to norepinephrine in rabbits: impact of endotoxin and fluid loading. Intensive Care Med
18. Booke M, Traber LD, Traber DL. Development of effective therapies for sepsis. Crit Care Med
19. VanderMeer TJ, Wang H, Fink MP. Endotoxemia causes ileal mucosal acidosis in the absence of mucosal hypoxia in a normodynamic porcine model of septic shock. Crit Care Med
20. Mitaka C, Hirata Y, Yokoyama K, Makita K, Imai T. A selective inhibitor for inducible nitric oxide synthase improves hypotension and lactic acidosis in canine endotoxic shock. Crit Care Med
21. Mitaka C, Hirata Y, Yokoyama K, Makita K, Imai T. S
-methylisothiourea sulfate improves renal, but not hepatic dysfunction in canine endotoxic shock model. Intensive Care Med
22. Shatney CH, Read G. Another unacceptable model of primate septic shock. Adv Shock Res
23. Halmagyi DF, Gillet DJ. Norepinephrine response in normal and hypotensive sheep. J Appl Physiol