Moderate-dose Vasopressin Therapy May Impair Gastric Mucosal Perfusion in Severe Sepsis: A Pilot Study

• Vasopressin has been suggested as a supplementary vasopressor in patients with vasodilatory shock.

• In septic states, high-dose vasopressin has been shown to exert negative effects on global hemodynamics.

• Administration of moderate vasopressin doses in severe sepsis allows norepinephrine dosages to be reduced and does not impair hepatosplanchnic perfusion. However, gastric mucosal perfusion is decreased.

SEPSIS is the 10th most common cause of death in industrialized countries, and is the leading cause of death in the intensive care unit.1Approximately 751,000 cases of severe sepsis occur annually in the United States.2A German survey in 2007 revealed that severe sepsis and/or septic shock occurred in 75,000 inhabitants (110 cases per 100,000 population) and sepsis in 79,000 inhabitants (116 cases per 100,000 population), therefore contributing to 60,000 deaths annually in Germany.3Maintenance of adequate tissue oxygen transport can be considered a primary objective in the management of patients with severe sepsis. Vasopressor therapy is a mainstay in achieving this goal. Unfortunately, global hemodynamic parameters, which represent the main target of vasopressor therapy, do not correlate with microvascular perfusion.4For this reason, assessment of the possible deleterious effects of vasopressor agents on tissue oxygenation at the microvascular level is essential.

Vasopressin has been used as a supplementary vasopressor in patients with advanced vasodilatory shock.5–9Vasopressin restores vascular tone in catecholamine-resistant shock states. This may be achieved by activation of vasopressin (V1) receptors, modulation of potassium-adenosine triphosphate channels, modulation of nitric oxide, and potentiation of adrenergic and other vasoconstrictor agents.10Because of its vasoconstrictive effects in the splanchnic area, vasopressin has been widely used for the management of gastrointestinal bleeding in patients with cirrhosis.11,12It has not yet been determined whether the effects of vasopressin on splanchnic blood flow may be more pronounced than those of other vasoconstrictors (e.g. , norepinephrine) in patients with severe sepsis and septic shock.

In a previous study,13we examined the effects of high-dose vasopressin on global hemodynamics and hepatosplanchnic blood flow in sepsis when norepinephrine was replaced completely by vasopressin. We found that hepatosplanchnic blood flow was preserved despite a substantial reduction in cardiac output, but that the difference between the gastric mucosal and arterial carbon dioxide tension [(Pco²-gap)] was increased, suggesting that gastric mucosal perfusion was impaired by high-dose vasopressin. Whether such potentially harmful effects also exist when vasopressin is used in lower doses has not yet been satisfactorily studied. It has been demonstrated that septic states cause an absolute or relative vasopressin deficiency.5,14However, it remains unclear whether vasopressin administered in lower doses to offset such a deficiency is associated with more beneficial effects than when it is used as a vasopressor.

The aim of our study was, therefore, to assess the effects of moderate doses of vasopressin on global hemodynamics, hepatosplanchnic blood flow, and gastric mucosal Pco².

The study was approved by our institutional research review committee (Ethik-Kommission der FSU Jena, Jena, Germany), and informed consent was received from each patient's next of kin before inclusion. Twelve consecutive patients were included in this pilot study within 24 h of development of severe sepsis, defined according to the definition of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Criteria.15 

All patients were treated with mechanical ventilation during the study period with an inspiratory:expiratory ratio of 1:1.5, a positive end-expiratory pressure of 6–10 cm H²O, and a constant inspired fraction of oxygen. Patients were sedated using continuous infusions of midazolam (maximum dose 9.0 mg/h) or propofol (maximum dose 300 mg/h) and sufentanil (maximum dose 75 μg/h). Ventilatory parameters and dose of analogosedation were kept constant during all measurements. All patients were treated with H²receptor antagonists (ranitidine), and enteral nutrition was stopped at least 2 h before the start of the study. All patients had arterial and central venous lines in place. According to our internal intensive care unit guidelines, all patients were treated with dobutamine (2.0–6.0 μg · kg⁻¹· min⁻¹). The infusion rate of dobutamine was not changed during the study period. The management of severe sepsis was in accordance with the guidelines of the German Sepsis Society and the German Society of Intensive Care.16 

Radial artery pressures and central venous pressures were measured with reference to the midaxillary line (transducer 5265 039, Viggo-Spectra-med, Bilthoven, The Netherlands). Cardiac output was measured using the transpulmonary thermodilution method using 15 ml of cooled (6–12 °C) saline solution and an arterial thermosensor in the femoral artery (PiCCO^TMsystem, Pulsion Medical Systems, Munich, Germany). Blood samples for blood gas estimations were drawn in duplicate, stored on ice, and processed within 30 min. Arterial and hepatovenous oxygen tension, pH and lactate concentration (ABL 3, Radiometer, Copenhagen, Denmark), and hemoglobin and oxygen saturation (Hemoximeter Osm 3, Radiometer) were measured immediately after sampling. Global oxygen consumption (VO²) was calculated using a metabolic monitor (Deltatrac II, Datex-Ohmeda, Helsinki, Finland). Urine output was monitored hourly throughout the study period.

Using continuous radiologic monitoring, a 7.5-F catheter was inserted into the hepatic vein via  the right internal jugular vein or the right femoral vein, and the position was verified radiologically. Hepatosplanchnic blood flow was evaluated by a continuous infusion of indocyanine green (ICG, Pulsion Medical Systems, Germany).17This technique enables measurement of total hepatosplanchnic venous blood flow. The dye used is a monosodium salt of ICG prepared in freeze-dried form and dissolved in sterile water. A 30-mg bolus of ICG was injected via  the central venous line and followed by a continuous ICG infusion (30 mg/h). To keep the volume of redistribution constant, no additional fluids were infused during the measurements. ICG was assayed spectrophotometrically at 805 nm on plasma samples. For each measurement, three samples were taken from the hepatic vein catheter and arterial line at an interval of 1 min and the results of the ICG measurements were averaged. Hepatosplanchnic blood flow was calculated using the equation18:

where SBF is the splanchnic blood flow (l/min), Cⁱis the ICG concentration of infusate (mg/l), and C^aand C^lvare the ICG concentrations in the radial artery and hepatic vein (mg/l), respectively. Fractional splanchnic blood flow (fSBF) was calculated using the equation:

Oxygen content was calculated as (hemoglobin [mg/dL]· 1.39 · hemoglobin oxygen saturation) + (arterial partial pressure of oxygen [mmHg]· 0.0031). Total body and splanchnic oxygen delivery (DO²) rates were calculated by multiplying the arterial oxygen content with the appropriate flow parameters. Splanchnic VO²was calculated by multiplying the hepatic venous oxygen-content difference with the appropriate splanchnic blood flow.

Gastric mucosal Pco²was determined by air tonometry (Tonometrics catheter, 16 F, Tonocap monitor, Datex-Ohmeda, Helsinki, Finland) every 10 min. The Pco²-gap was calculated as the difference between gastric mucosal Pco²and Pco²in the corresponding arterial blood sample.

Before data collection was started, all patients were stabilized by volume loading, predominantly using a balanced electrolyte solution (Jonosteril®, Fresenius Kabi, Germany), until there was no further increase in cardiac output and targeting a central venous pressure of 8 mmHg. The norepinephrine dose was increased to achieve a mean arterial pressure of at least 70 mmHg. Measurements began when the patients were stable (no additional volume demand and an unchanged norepinephrine dose for 4 h) to avoid possible confounding effects of changes in therapy and fluctuation in hemodynamic parameters (fig. 1). After baseline measurements were obtained, arginine vasopressin (0.04 IU · kg⁻¹· h⁻¹) was administered and norepinephrine dose was reduced in increments of 10% to achieve the baseline mean arterial pressure. After 4 h all measurements were repeated, and the vasopressin infusion was terminated. All measurements were taken with the patients in the supine position, and no interventions were performed during this period apart from adjustment of vasopressor dosage according to the study protocol.

All values are presented as median [interquartile range]. Paired data obtained before and 4 h after vasopressin infusion were compared with the Wilcoxon signed rank nonparametric test (Medcalc®, version 4.16e, Medcalc Software, Mariakerke, Belgium). All statistical tests were two-tailed and statistical significance was considered at P < 0.05.

Twelve patients with severe sepsis were included in a consecutive series design study. Patients were between 40 and 89 yr old with an Acute Physiology and Chronic Health Evaluation (APACHE)-II score between 18 and 36 on admission to the intensive care unit. Patient demographics, including norepinephrine and dobutamine dosages, are summarized in table 1. Administration of vasopressin (0.04 IU · kg⁻¹· h⁻¹) enabled norepinephrine dosage to be significantly reduced, from a median of 0.23 [0.19–0.85] to 0.11 [0.08–0.36]μg · kg⁻¹· min⁻¹(P = 0.002), whereas the mean arterial pressure remained constant (table 2). No additional fluids were given during the study period apart from the basic crystalloid infusions (range: 80–250 ml within 4 h).

Infusion of vasopressin and reduction of norepinephrine caused minimal changes in global hemodynamics (table 2). Heart rate decreased significantly from 99 [81–115] to 96 [74–109] min⁻¹(P = 0.016). Cardiac index decreased significantly from 3.7 (2.8–4.7) L · min⁻¹· m⁻²to 3.5 (2.7–3.6) L · min⁻¹· m⁻²(P = 0.003) (table 2and fig. 2). Intrathoracic blood volume index and central venous pressure remained unchanged (table 2). In parallel with cardiac output, global DO²index decreased significantly from 461 (375–637) to 419 (352–551) ml · min⁻¹· m⁻²(P = 0.002), whereas global VO²index did not change after vasopressin infusion.

Total hepatosplanchnic blood flow (fig. 3), fractional splanchnic blood flow as part of cardiac output, splanchnic DO², splanchnic VO², and hepatic venous oxygen saturation did not change significantly after vasopressin infusion (table 2). Hepatosplanchnic oxygen extraction (fig. 4) and hepatic ICG extraction (fig. 5) showed marked interindividual variability after vasopressin infusion but remained globally unchanged. However, gastric mucosal Pco²-gap increased significantly from 13.3 [8.0–16.7] to 17.1 [10.3–28.7] mmHg (P = 0.002) (fig. 6). Only four patients received propofol in our study (table 1). Gastric mucosal Pco²-gap increased after administration of vasopressin in patients who were sedated using midazolam (14 [7.2–16.7] to 17.1 [9–27.9] mmHg, P = 0.028) and in those who were sedated using propofol (11.1 [6.7–24.5] to 20.9 [9.7–38.4] mmHg, P = 0.066).

Arterial lactate concentration (1.9 [1.4–2.7] to 2.0 [1.5–2.3] mM), hepatovenous lactate concentration (1.6 [0.9–1.2] to 1.7 [1.0–1.3] mM), and hourly urine production (74 [57–130]versus  107 [60–180] ml/h) did not change significantly during or after administration of vasopressin.

The major finding of our study is that administration of a moderate so-called substitution dose of vasopressin in severe sepsis enabled norepinephrine dosages to be reduced, but did not reduce total hepatosplanchnic perfusion and minimally decreased cardiac output and global DO². However, gastric mucosal perfusion decreased markedly as evidenced by an increase in the gastric mucosal Pco²-gap after vasopressin infusion.

In a previous study, we showed that use of vasopressin as a vasopressor in higher dosages was associated with negative effects on global hemodynamics.13In the current study, using a lower dosage of vasopressin, global hemodynamic effects were minimal, including a slight decrease in heart rate and cardiac index. In a state of relative vasopressin deficiency, vasopressin acts as a vasopressor; however, it has no positive inotropic or chronotropic properties, in contrast with norepinephrine in higher dosages.13The reduction in cardiac index and heart rate may, therefore, be due to the reduction in the norepinephrine dosage after the vasopressin infusion. The reduction in global DO²observed in our study had no significant effect on global VO²or arterial lactate concentration.

In experimental studies, vasopressin has been convincingly shown to have vasoconstrictive effects on gut perfusion. In pigs, administration of vasopressin was associated with a significant decrease in blood flow in the superior mesenteric artery after cardiopulmonary resuscitation.19In our study, however, vasopressin infusion at a dosage of 0.04 IU · kg⁻¹· h⁻¹had no effect on total hepatosplanchnic blood flow or fractional hepatosplanchnic blood flow, similar to our previous observation using higher vasopressin dosages.13Nevertheless, we used a liver vein catheter for measurement of hepatosplanchnic blood flow, a technique that only enables calculation of the net flow through the liver and does not allow differentiation between mucosal, hepatoarterial, and portal venous flows. Despite the relatively invasive nature of this procedure, it is considered the only method by which hepatosplanchnic flow can be measured in clinical practice without surgical intervention.20In our study it was important to differentiate the possible influence of vasopressin administration on tissue perfusion at the global, regional, and local mucosal levels. Therefore, measurement of the regional heptosplanchnic flow was essential.

Importantly, we observed a significant increase in gastric mucosal Pco²-gap in our patients after vasopressin infusion, denoting impaired gastric mucosal perfusion. These data are in accordance with our previous observation using a higher vasopressin dosage.13Intestinal and gastric mucosal vasoconstriction were recently demonstrated in cardiac surgery patients for a dose range of 1.2–4.8 IU vasopressin per hour.21Impaired gut mucosal perfusion has also been demonstrated in septic rats.22However, Dünser et al.  23and Patel et al.  24reported no increase in gastric mucosal carbon dioxide production using similar vasopressin dosages. The differences with our findings can perhaps be explained by the increased mean arterial pressure23or the concomitant use of dopamine24in contrast with dobutamine that was used in our study. Our own results support those of van Haren et al. ,25who demonstrated that a so-called substitution dose of 0.04 IU/min led to supranormal plasma vasopressin levels of more than 200 pg/ml, and that there was a correlation between plasma vasopressin levels and mucosal Pco²-gap.

The increase in gastric mucosal Pco²-gap after vasopressin infusion in our patients despite the preserved heptosplanchnic flow may be explained by redistribution of flow. Our results also emphasize the heterogeneity of tissue perfusion in patients with severe sepsis, previously reported using other techniques.26It is also important to stress that these alterations were independent of global hemodynamic measurements, confirming the results of previous reports.4,26–28 

Gastric mucosal tonometry is one of the only techniques that can be used at the bedside to assess gastrointestinal mucosal perfusion, but it has important limitations, and whether it can be a useful tool in the intensive care unit is still unclear.29,30Lang et al.  reported in a rat model that changes in blood flow with sepsis varied among the different splanchnic organs.31In stable patients with sepsis, the Pco²-gap was not correlated with global measures of hepatosplanchnic oxygenation, including hepatosplanchnic blood flow, the suprahepatic venous blood oxygen saturation, and the mesenteric venoarterial Pco²gradient.17However, Kavarana et al.  demonstrated that gastric hypercarbia was associated with adverse postoperative morbidity.32 

Although we found that vasopressin infusion was associated with impaired gastric mucosal perfusion, the possible effect of these alterations on outcome is yet to be determined. Indeed, a large multicenter randomized controlled study reported that low-dose vasopressin was not associated with reduced mortality rates compared with norepinephrine among patients with septic shock.33The authors demonstrated a potential benefit of vasopressin in patients with low to moderate norepinephrine dosage.33A subsequent subgroup analysis showed improved outcome in septic patients who were treated with a combination of vasopressin and corticosteroids.34However, in this study,33vasopressin was administered at a lower dose than that used in our study.

Our study has some limitations. First, the sample size is small, which increases the likelihood of type II error. Our study should, therefore, be considered as a pilot study, and additional studies with adequately powered sample sizes are needed to confirm our observation. Second, vasopressin was administered for only 4 h. We cannot elaborate, therefore, on whether the effects of vasopressin infusion on gastric mucosal microvascular alterations may persist if used for longer periods. Third, our results cannot be extrapolated to lower dosages of vasopressin. Fourth, different sedation regimens were used, which may have had different effects on our study parameters. Nevertheless, the choice of sedating agents was left to the discretion of the attending physician to avoid inappropriate prolongation of mechanical ventilation due to the use of agents with relatively long half-lives. However, the dosage of sedation was kept constant during all measurements and the alterations in gastric mucosal perfusion occurred in patients who received midazolam and those who received propofol.

Vasopressin at a dosage of 0.04 IU · kg⁻¹· h⁻¹may impair gastric mucosal perfusion with minimal global and regional hemodynamic effects. The routine use of vasopressin in patients with severe sepsis may, therefore, be associated with potential risks.

The authors thank Mrs. Ilona Witte and Mrs. Astrid Trinks (Technicians, Department of Anesthesiology and Intensive Care, Friedrich Schiller University, Jena, Germany) for medical-technical assistance.