Optimal Adrenergic Support in Septic Shock Due to Peritonitis

This study was approved by our Institutional Review Board for animal care (Brussels, Belgium). Care and handling of the animals were in accord with National Institutes of Health guidelines. Twenty-eight mature, female sheep (weight, 30.5 ± 3.6 kg) were included in the study. After tracheal intubation under intramuscular injection of 2 ml xylazine (Bayer, Leverkusen, Germany) and 3 ml ketamine (Ketalar®; Warner-Lambert Manufacturing Ltd., Dublin, Ireland), each sheep was anesthetized with infusion of a mixture of midazolam (Dormicum®; Hoffmann-La Roche, Basel, Switzerland) and fentanyl (Janssen Pharmaceutica, Beerse, Belgium) using an infusion pump (Perfusor® secura; Braun, Melsungen AG, Germany) and mechanically ventilated with an inspired oxygen fraction (Fio²) of 0.5 (Servo ventilator 900B; Siemens-Elema, Solna, Sweden). Muscle paralysis was obtained by the administration of pancuronium bromide, administered at an initial dose of 0.15 mg/kg and subsequent infusion of 0.075 mg · kg⁻¹· h⁻¹. Respiratory rate was 14 breaths/min, and tidal volume was adapted to keep end-tidal partial pressure of carbon dioxide (Pco²; 47210A Capnometer; Hewlett-Packard, Waltham, MA) between 28 and 38 mmHg. The left forepaw vein was catheterized for the intravenous administration of anesthetic agent and pancuronium bromide. The right forepaw vein was catheterized for intravenous infusion of lactated Ringer's solution. The right femoral artery was catheterized for monitoring of arterial blood pressure and withdrawal of arterial blood samples. Through the right jugular vein, a balloon-tip pulmonary artery catheter (93A-439H-7.5F; Baxter Edwards Critical-Care, Irvine, CA) was placed under guidance of pressure waves (Sirecust™ monitor 404; Siemens, Davis, CA).

Through a midline laparotomy, the cecal and ileocecal junction was identified. After a 1-cm perforation in the cecal tip, fecal material (approximately 30 ml) was spilled into the peritoneal cavity toward the right lower quadrant. Ultrasonic flow probes (Transonic® Flowprobe; Transonic Systems Inc., Ithaca, NY) were placed around the left femoral artery and the superior mesenteric artery for simultaneous determinations of regional blood flow. A tonometric catheter (TRIP, NGS catheter; Tonometrics, Helsinki, Finland) was inserted into the ileum to measure ileal intramucosal carbon dioxide tension 20,21(in a pilot study, gastric tonometry was found to be unreliable in these ruminant animals), and the abdomen was closed with a running suture of 0 Dexon® (Davis & Geck, Wayne, NJ). During the whole experiment, feces were not removed, and the peritoneal cavity was not washed. A Foley catheter was introduced via  the urethra to collect urine.

In each group, fluid maintenance was approximately 1,000 ml lactated Ringer's solution during the course of surgery, titrated to keep the pulmonary artery occlusion pressure (PAOP) constant as at the first measurement. Body temperature was kept below 41°C by ice package. In pilot experiments, the polymicrobial peritonitis and bacteremia were confirmed by blood cultures.Organisms most frequently grown included Klebsiella oxytoca , Enterobacter cloaceae , Enterococcus faecalis , Citrobacter freundii  and various strains of Escherichia coli .

The sheep were randomly allocated to one of the four experimental groups of seven sheep each.

No adrenergic agent was administered.

When MAP decreased to a value less than 75 mmHg, a norepinephrine infusion (Levophed®; Sanofi-Pharma, New York, NY; 8 mg/4 ml) was started at an initial rate of 0.5 μg · kg⁻¹· min⁻¹and increased by 0.5-μg · kg⁻¹· min⁻¹increments up to a maximum of 5 μg · kg⁻¹· min⁻¹to maintain MAP between 75 and 85 mmHg.

When MAP decreased to a value less than 75 mmHg, a dopamine (Dynatra 200; Sintesa, Brussels, Belgium; 200 mg/5 ml) infusion was started and titrated to maintain MAP between 75 and 85 mmHg. The initial rate was 5 μg · kg⁻¹· min⁻¹and increased by 2.5-μg · kg⁻¹· min⁻¹increments up to a maximum rate of 20 μg · kg⁻¹· min⁻¹. If MAP remained lower than 75 mmHg with a dose of 20 μg · kg⁻¹· min⁻¹, a norepinephrine infusion was added at an initial concentration of 0.5 μg · kg⁻¹· min⁻¹and increased up to 5 μg · kg⁻¹· min⁻¹, as in group 2.

Treatment was identical to that of group 2 except that dobutamine (Dobutrex®; Eli Lilly, Indianapolis, IN; 250 mg/ 20 ml) was added at the same time as norepinephrine and titrated to obtain a 15% increase in cardiac output. The infusion rate was initially 5 μg · kg⁻¹· min⁻¹and increased by 2.5-μg · kg⁻¹· min⁻¹increments up to a maximum of 20 μg · kg⁻¹· min⁻¹.

In each sheep, the administration of lactated Ringer's solution was titrated to keep PAOP as close as possible to the baseline value, and the doses of adrenergic agents were titrated at 5-min intervals. Potassium chloride was added to lactated Ringer's solution as needed to keep plasma potassium concentration between 3.5 and 5.0 mm.

One hour was allowed before the first measurements (as baseline values) were taken. Measurements were repeated every hour throughout the experiment and included MAP, pulmonary arterial pressure, PAOP, right atrial pressure, cardiac output, femoral and superior mesenteric arterial blood flows, expired oxygen fraction, end-tidal Pco², minute volume, blood gases, arterial hemoglobin concentration, blood arterial lactate concentration, plasma electrolytes concentration, infusion volume, and urinary output. The survival time was also recorded.

Pressures were monitored continuously, using a pressure monitoring kit (Baxter, Uden, Holland) with amplifiers (Servomed; Hellige, Freiburg, Germany) and a pen recorder (2600S; Gould, Instruments Division, Cleveland, OH). All pressures were determined at end-expiration. Cardiac output was measured in triplicate by the thermodilution technique (Swan-Ganz Catheter; Baxter, Irvine, CA), using 10-ml iced saline solution (0°C) at end-expiration. Stroke volume (SV), SVR, pulmonary vascular resistance (PVR), and left ventricular stroke work (LVSW) were calculated by standard formula. Superior mesenteric and femoral artery blood flows were simultaneously measured by ultrasound volume flowmeter (T208; Transonic Systems Inc.; calibrated by the manufacturer).

Exhaled gases were passed directly through a mixing chamber for sampling of expired oxygen fractions (P.K. Morgan Ltd., Chatham, Kent, United Kingdom). Expired minute volume was measured with a spirometer (Haloscale Respirometer; Wright, Edronton, London, United Kingdom) over a 2-min period. Arterial and mixed venous blood samples were simultaneously withdrawn for immediate determination of blood gases (ABL625; Radiometer, Copenhagen, Denmark), arterial and mixed venous oxygen saturations, and total hemoglobin (OSM 3 Hemoximeter; Radiometer). Blood lactate and electrolyte (K⁺, Na⁺, Ca²⁺, Cl⁻) concentrations were determined by an automated analyzer (ABL625; Radiometer). Ptco²was measured by saline tonometry using the standard technique. The tonometer balloon was filled with 2.5 ml saline and allowed to equilibrate for approximately 60 min. Saline was anaerobically aspirated, the first milliliter was discarded, and the remaining 1.5 ml was analyzed immediately using the blood gas analyzer. The Pco²gap was calculated as the difference between Ptco²and arterial Pco². Do², V̇o², and oxygen extraction ratio were derived from standard equations.

When the sheep died, tissue samples were surgically excised from the lung, the liver, the gut, and the kidney and immediately immersed in 4% formalin (pH = 7). The samples were sectioned and stained with hematoxylin and eosin for light microscopy. All biopsies were coded to enable processing and evaluation. A pathologist unfamiliar with the conditions of the study examined all fields of the slides in a blinded fashion, and 10 micrographs from each specimen were taken from consecutive squares. Anatomic abnormalities were recorded semiquantitatively as focal cell injury, edema, and acute inflammation. The degree of each abnormality was graded numerically from 0 (normal) to 3 (most severe) in the intestine specimens and from 0 (normal) to 2 (most severe) in the other specimens. 22 

Data were analyzed using a two-way analysis of variance for repeated measurements followed by a Newman-Keuls procedure. Missing values were estimated using the method proposed by Yates that minimizes the error sum of squares. 23The error degree of freedom is reduced by the number of missing data. By doing this, (1) the estimates of group and time effects are exactly the same as those obtained by the correct least square procedure, and (2) the error sum of squares is exactly the same as given by the correct procedure. The correct least square approach used the type III or type IV sum of squares in which the characteristics equations are solved by the general inverse technique. A Kruskal-Wallis test was used to evaluate the differences in the degrees of anatomic abnormalities. To evaluate survival time, Kaplan-Meier survival curves were constructed and analyzed by the Mantel-Cox log-rank test. P < 0.05 was considered statistically significant. All values are expressed as mean ± SD, except when mentioned otherwise.

No significant differences were observed between the four groups in baseline hemodynamic and metabolic parameters (table 1). All animals developed severe hyperthermia requiring ice cooling. The control group was characterized by arterial hypotension, tachycardia, increased cardiac output, and low SVR. The mean amount of fluid administered was 10–15 ml · kg⁻¹· h⁻¹; there were no statistical differences among groups in the amount of fluid administered. Blood cultures were performed consecutively in the first three sheep and showed positive and consistent results in all three animals. The start time of adrenergic treatment after the end of the surgical procedure was 9.7 ± 2.1 h in the norepinephrine group, 10.0 ± 2.3 h in the dopamine-norepinephrine group, and 9.9 ± 2.4 h in the dobutamine-norepinephrine group, respectively (differences not significant). In all sheep in the dopamine-norepinephrine group, norepinephrine had to be added to dopamine within 2 h of dopamine administration to maintain MAP.

Mean arterial pressure was well maintained for at least 15 h in all treatment groups (fig. 1). Cardiac output increased in all groups, but this increase was maintained only in the dobutamine-norepinephrine group (fig. 1), partly because of an increase in heart rate. LVSW transiently increased in all treatment groups, and this increase was prolonged in the dobutamine-norepinephrine group. SVR increased in the dopamine-norepinephrine and norepinephrine groups but decreased in the dobutamine-norepinephrine group (table 1). Superior mesenteric blood flow increased in all groups, but this increase was maintained only in the dobutamine-norepinephrine group (fig. 2). Do²increased significantly in all treatment groups, especially in the dobutamine-norepinephrine group. Toward the end of the experiment, Do²and V̇o²were greater in the dobutamine-norepinephrine group than in the dopamine-norepinephrine or norepinephrine alone group (table 1).

Blood lactate concentration and Pco²gap increased less in the dobutamine-norepinephrine group than in the control group (fig. 3). Cumulative urine output was significantly higher in the dobutamine-norepinephrine group than in the other groups (fig. 4).

Three of seven sheep survived longer than 24 h in the dobutamine-norepinephrine group, two of seven survived longer than 24 h in the dopamine-norepinephrine group, and none survived longer than 24 h in the norepinephrine or control groups. The mean survival time was 17 ± 2 h in the control group, 20 ± 1 h in the norepinephrine group, 24 ± 6 h in the dopamine-norepinephrine group, and 24 ± 5 h in the dobutamine-norepinephrine group. The survival time was significantly longer in all treatment groups than in the control group (P < 0.05;fig. 5). The survival time was also significantly longer in the dobutamine-norepinephrine and dopamine-norepinephrine groups than in the norepinephrine alone group (P < 0.05).

Results of anatomic examination of the organs are shown in figure 6. Lung biopsies revealed variable degrees of inflammatory change from interstitial to alveolar edema and congestive atelectasis to severe injuries, including diffuse alveolar damage with exudation in the interstitium and alveolar spaces, and hemorrhagic alveolitis. Inflammation and edema formation were less severe in the dopamine-norepinephrine and dobutamine-norepinephrine groups (P < 0.02 for all groups). Liver alterations were characterized by congestion, sinusoidal pericentrolobular dilatation, and necrosis. These alterations were least severe in the dobutamine-norepinephrine group (P < 0.05 for all groups). Small intestine congestion and edema were also least severe in the dobutamine-norepinephrine group than in the other groups (P < 0.01 for all groups). Kidney examination showed only congestion. Figure 7shows light micrographs of pulmonary, hepatic, and renal biopsies from one sheep in the control group.

The current peritonitis model produced by the spillage of feces followed by generous fluid administration reproduces many of the clinical features of human septic shock, including hyperthermia, arterial hypotension, tachycardia, increased cardiac output, low SVR, and hyperlactatemia. The model is characterized by large fluid requirements (10–15 ml · kg⁻¹· h⁻¹) to achieve this hyperkinetic state. In the absence of antibiotic administration and with no other support than intravenous fluid loading, the model is lethal after 15–20 h. Although we acknowledge that the use of anesthesia may have influenced our results, ketamine is the anesthetic agent with the fewest cardiovascular effects in septic shock. 24Recent experimental studies have suggested a gender advantage for proestrus female mice over males in survival from sepsis, 25and ovariectomized mice had a higher death rate than proestrus mice after cecal ligation and puncture. 26These potential difference were limited in the current study by using only female sheep, and although the estrus state of individual sheep was not measured, the randomization of the sheep into the four study groups should have limited any potential influence of varying concentrations of female hormones on the response to sepsis.

Dopamine is often recommended as the first-line agent to correct fluid resistant hypotension, 1,2but its use has been challenged recently on the basis that the dopaminergic effects have not been shown to be beneficial. 27It has also been argued that dopamine is not very effective, so that in many patients, an adequate tissue perfusion pressure cannot be obtained with the use of dopamine alone, even at doses as high as 80 μg · kg⁻¹· min⁻¹. 28,29In a porcine endotoxic shock model, Sautner et al.  30found that the administration of norepinephrine did not improve endotoxin-induced damage of the ileum and colon mucosa. However, low-dose (3 μg · kg⁻¹· min⁻¹) dopamine administration delayed and attenuated the decrease in intestinal villus blood during normotensive endotoxemia in the rat. 31In the current study, dopamine alone was insufficient to maintain cardiac output or MAP, and norepinephrine had to be added within 2 h after the onset of dopamine infusion. By its stronger α-adrenergic effect on vascular tone, norepinephrine effectively corrected hypotension. Nevertheless, cardiac output and LVSW increased with dopamine, and MAP was better maintained by the combination of dopamine with norepinephrine than by norepinephrine alone.

Dobutamine is the catecholamine of choice to increase myocardial contractility, cardiac output, and Do². 13,32–35The addition of dobutamine to vasopressor regimens can markedly increase Do²and Vo²in septic patients. 36,37Martin et al.  38found that the addition of dobutamine to norepinephrine in septic shock patients significantly improved MAP, cardiac output, SV, and LVSW. In a rat model, dobutamine prevented intestinal arteriolar constriction and maintained villus blood flow at preendotoxemic values, 39and in a porcine model of endotoxic shock, dobutamine also increased blood flow to the gastrointestinal mucosa. 40The vasodilating effect of dobutamine on gastric mucosal microcirculation has also been noted in humans. 41In our study, addition of dobutamine to norepinephrine increased cardiac output and resulted in higher MAP, SV, cardiac output, and LVSW when compared with the combination of dopamine with norepinephrine or norepinephrine alone during the final stages of the experiment.

Superior mesenteric blood flow was better maintained in the dobutamine-norepinephrine group than in the other groups, reflecting an improvement in splanchnic perfusion. Several studies have shown that dobutamine can decrease Pco²gap and increase pHⁱ. 16,41Although the use of the tonometer in ruminant animals may be problematic, the ileal Pco²gap increased less in the dobutamine-norepinephrine group than in the other groups. Blood lactate concentrations were significantly lower in the dobutamine-norepinephrine group than in the control group. Even though the increase in lactate concentrations in sepsis can result from many complex cellular mechanisms other than just tissue hypoxia, including catecholamine-stimulated aerobic glycolysis, 42the earlier reduction in blood lactate concentrations suggests a more effective improvement in cellular metabolism and is associated with improved outcome. 43 

Our model also allowed us to study the effects of adrenergic support on survival. Lower mortality rates have been reported for patients who reached supranormal values of Do²with adrenergic support. 15,44Recently, Rivers et al.  45showed that aggressive resuscitation regimens targeting venous oxygen saturation increased survival rates in severe sepsis and septic shock. However, detrimental effects have also been observed when therapy is oriented to supranormal Do²and V̇o²levels in critically ill patients. 19In the current study, the combination of norepinephrine with dobutamine or dopamine prolonged survival.

Globally, the dobutamine-norepinephrine group showed the least anatomic damage to the organs, suggesting the beneficial effects of a higher Do²to these organs. In a porcine model of fecal peritonitis, Tighe et al.  46,47found that dobutamine had detrimental effects on the histology in the liver, with marked hepatocellular damages. Our current study does not support these findings, perhaps because our peritonitis model is less acute, fluid administration was more generous, and the dose of dobutamine was limited to a maximum of 20 μg · kg⁻¹· min⁻¹as currently recommended. 1 

In this hyperdynamic septic shock model, we conclude that, when volume replacement is adequate and perfusion pressure is well maintained by vasopressor therapy, dopamine alone is insufficient to maintain cardiac output or MAP. The combination of norepinephrine with either dopamine or dobutamine is more efficient than norepinephrine alone in reversing the hemodynamic abnormalities, alleviating the histologic injury in the major organs, and prolonging survival. In particular, the combination of norepinephrine with dobutamine can limit the increase in arterial lactate concentration and Pco²gap and attenuate anatomic damage to the lung, liver, and small intestine. These data caution against the use of norepinephrine alone in the treatment of hypotension in septic shock.