The infusion of low-dose dopamine (defined as a dose lower than 5 μg/kg per min administered to normotensive patients) may not improve gut mucosal perfusion. In fact, Nevière and coworkers 27 showed that low-dose dopamine decreased gut mucosal blood flow in septic patients. Furthermore, other investigators 27282930 reported that either pHi or PCO₂ gap were unchanged in patients with sepsis treated with low-dose dopamine. The effects on liver blood flow may also be variable; Maynard and coworkers 30 observed that dopamine did not affect ICG clearance and monoethylglycinexylidide (MEGX) formation from lidocaine. Interestingly, the effects of dopamine on splanchnic blood flow may differ according to basal splanchnic perfusion. Low-dose dopamine increased splanchnic blood flow that was low at baseline (seven patients) but not when splanchnic perfusion was preserved (four patients) 28. The very small number of patients in each group limited these observations. Recently, Jakob and coworkers 31 reported that dopamine administration titrated to achieve a 25% increase in cardiac output induced a significant increase in splanchnic blood flow from 0.9 to 1.1 l/min per m², which was associated with a significant reduction in splanchnic oxygen consumption.

The results are even more controversial when dopamine is used at higher doses to restore blood pressure. Ruokonen and coworkers 32 observed that dopamine increased splanchnic blood flow and metabolism in some but not all patients with septic shock. In some patients, the same group of investigators 33 also observed an increase in hepatic vein oxygen saturation, suggesting an improvement in the balance between oxygen supply and demand during dopamine administration. However, in a pilot study, Marik and Mohedin 34 reported that dopamine administered at doses up to 25 μg/kg per min even decreased pHi. Given the very small number of patients included in these studies, no definite conclusions can be drawn regarding the effects of dopamine on splanchnic blood flow in septic patients.

Comparison of the effects of norepinephrine and dopamine is difficult because norepinephrine is often combined with dobutamine, and study results are conflicting. Ruokonen and coworkers 32 reported unpredictable effects on splanchnic blood flow in patients with septic shock with norepinephrine, whereas dopamine induced a consistent increase in splanchnic blood flow. By contrast, in the randomized study reported by Marik and Mohedin 34, conducted in 20 septic patients with hyperdynamic septic shock, dopamine was reported to induce a decrease in pHi when compared with norepinephrine. More recently, De Backer and coworkers 35 reported the effects of dopamine, norepinephrine and epinephrine on the splanchnic circulation in moderate and in severe septic shock, and the main results are as follows. In moderate septic shock cardiac index was similar in dopamine-treated and norepinephrine-treated patients, and higher in epinephrine-treated patients, whereas splanchnic blood flow was the same with the three drugs. The gradient between mixed venous and hepatic venous oxygen saturation gradient was the lowest with dopamine, while PCO₂ gaps were identical. In patients with moresevere septic shock cardiac index was greater and splanchnic blood flow lower with epinephrine than with dopamine and epinephrine; mixed venous and hepatic venous oxygen saturation gradient was greater with epinephrine, whereas PCO₂ gap remained unaltered by any of the treatments.

Given the available data (summarized in Table 1), no definite conclusions can be drawn regarding differences between dopamine and norepinephrine on splanchnic blood flow and metabolism in patients with septic shock.

In patients with sepsis, a retrospective study conducted by Silverman and coworkers 24 identified a beneficial effect of dobutamine infusion on pHi 24. Two years later Gutierrez and coworkers 36 reported an increase in pHi with dobutamine infusion in patients with sepsis syndrome who initially had low pHi. This beneficial effect, confirmed in other studies 373839, was not related to an increase in splanchnic blood flow induced by dobutamine 3940. Creteur and colleagues 41 reported that dobutamine decreased the PCO₂ gap in septic patients with a high gradient between the mixed venous and hepatic vein oxygen saturation (>20%), whereas PCO₂ gap was not affected in patients when this gradient was less than 20%. This suggests that patients with the most severe alterations in hepatosplanchnic blood flow are also prone to decreased mucosal perfusion.

Dobutamine usually, but not without exception, increases splanchnic perfusion 404142. The effects on splanchnic metabolism are more variable 39 and may depend on the adequacy of splanchnic perfusion at baseline. In patients with septic shock, De Backer and coworkers 43 reported that splanchnic oxygen consumption increased during dobutamine administration only in patients with an increased gradient between hepatic venous and mixed venous oxygen saturation.

Combinations of dobutamine and other catecholamines have often been studied, in particular in association with norepinephrine for its effects on β-receptors, with the aim of modulating hepatosplanchnic haemodynamics. Indeed, in patients with sepsis, changing from norepinephrine (α-agonist and β-agonist) to phenylephrine (pure α-agonist), titrated to produce similar global haemodynamic measurements, led to a decrease in splanchnic blood flow, splanchnic oxygen delivery and gastric pHi. These changes were associated with decreased rates of liver lactate uptake and glucose production 44.

Whether dobutamine has a specific effect on the splanchnic circulation is still debated. In a cross-over study conducted in eight patients with septic shock, Meier-Hellmann and coworkers 45 showed that epinephrine caused lower splanchnic flow and oxygen uptake, lower gastric pHi, and higher hepatic vein lactate concentration than did the combination of dobutamine and norepinephrine. Duranteau and coworkers 11 compared the effects of epinephrine, norepinephrine and the combination of norepinephrine and dobutamine in patients with septic shock on gastric mucosal flow, as assessed using a laser Doppler technique. Epinephrine and dobutamine–norepinephrine led to a significant increase in gastric mucosal flow as compared with norepinephrine alone, but these findings were not corroborated by those reported by Seguin and coworkers 46. Moreover, in patients with septic shock resistant to dopamine, the combination of norepinephrine and dobutamine, in comparison with epinephrine alone, restored gastric pHi more quickly and limited the increase in arterial lactate concentration. However, there was no difference in gastric mucosal PCO₂ gradients between groups at 24 hours of treatment 38.

The preferential effect of dobutamine on splanchnic blood flow was not confirmed by Reinelt and coworkers 42, who studied the effects of dobutamine on fractional splanchnic flow and hepatic glucose production in septic patients resuscitated adequately with fluid and norepinephrine. Their results showed a parallel increase in splanchnic blood flow and cardiac index, unaltered splanchnic oxygen consumption and decreased rate of endogenous production of hepatic glucose. These findings suggest that splanchnic blood flow is increased in well resuscitated septic patients, and that a dobutamine test is able to reveal a oxygen delivery/consumption dependency 4143 but it cannot exclude intraorgan blood flow redistribution at the microcirculatory level. The inadequacy of blood flow distribution is mirrored by the absence of correlation between splanchnic blood flow and the PCO₂ gap.

Reported data on the effects of dobutamine and norepinephrine on splanchnic haemodynamics are summarized in Tables 2 and 3, respectively.

We suggest that both dopamine and norepinephrine can be given to septic shock patients as first-line catecholamine drugs but that their use must be weighed against the undesired neuroendocrine side effects of dopamine 45. Epinephrine should be reserved for use as rescue therapy. If norepinephrine is chosen as the first agent, then the addition of dobutamine may be considered.

Physiologically, vasopressin (a nonapeptide that is released from the neurohypophysis) plays a minor role in blood pressure regulation. Clinical data revealed that the initially very high plasma concentrations of vasopressin decrease during prolonged sepsis 52.

In the past few years clinical studies showed that blood pressure can be rapidly restored in septic shock using vasopressin, but this is mainly at the expense of cardiac output 53. Nevertheless, in 2000 the American Heart Association and International Liaison Committee on Resuscitation recommended (grade IIB) continuous vasopressin infusion in refractory septic shock 54. However, the effects of vasopressin on regional (i.e. splanchnic) blood flow are discussed controversially.

In 1997, Landry and coworkers 52 reported on the continuous infusion of vasopressin (1.8–3.0 IU/hours) in five patients with septic shock. In all patients, blood pressure was rapidly restored and urine output increased in three. Patel and coworkers 55 randomly assigned 24 patients with septic shock to a double-blind 4-hour infusion of norepinephrine or vasopressin, and open-label vasopressors were titrated to maintain blood pressure. Although norepinephrine dosage could be significantly lowered in the vasopressin group, blood pressure and cardiac index were maintained in both groups. Urine output did not change in the norepinephrine group but increased substantially in the vasopressin group. Similarly, creatinine clearance did not change in the norepinephrine group but increased by 75% in the vasopressin group. Finally, gastric mucosal PCO₂ gradient did not change significantly in either group.

Recent results from Klinzing and coworkers 56, however, indicate that vasopressin may lead to a different blood flow distribution pattern in the splanchnic area as compared with norepinephrine. In 12 patients with septic shock, vasopressin was administered at a dose of 0.06–1.8 IU/min to replace norepinephrine completely. As a result, cardiac index and systemic oxygen uptake decreased significantly. Total splanchnic blood flow tended to decrease, while splanchnic blood flow expressed as percentage of cardiac output as well as the PCO₂ gap were doubled 56. By contrast, the increase in gastric PCO₂ gap suggests that blood flow may have been redistributed away from the mucosa, and therefore it does not appear beneficial to directly replace norepinephrine with vasopressin in septic shock. Clinical data also suggest that low-dose vasopressin (0.04 IU/min) to compensate for endogenous deficiency could be a beneficial strategy 57585960, as was recently demonstrated by Dünser and coworkers 61, who randomly assigned 48 patients with catecholamine-resistant vasodilatory shock to receive a combined infusion of vasopressin and norepinephrine or norepinephrine alone. Vasopressin-treated patients had significantly lower heart rate, norepinephrine requirement and incidence of new onset tachyarrhythmias. Mean arterial pressure, cardiac index and stroke volume were significantly greater, and the PCO₂ gap was significantly lower in patients treated with this combination. However, these patients also presented with a significant increase in plasma bilirubin concentration, suggesting an impaired liver blood flow and/or a depressed hepatic function mediated by vasopressin.

More recently, terlipressin (glycinpressin), a long-acting vasopressin analogue, was proposed as a treatment for septic shock. O'Brien and coworkers 62 reported their clinical experience with terlipressin (1–2 mg) as rescue treatment in eight patients with refractory septic shock. Those investigators reported a rapid and 24 hour lasting stabilization in blood pressure, with a significant reduction in norepinephrine but a significant decrease in cardiac index. In that study, seven patients required renal replacement therapy and four patients died during their stay in the ICU. However, optimism regarding these findings must be tempered somewhat 63, in particular because detrimental effects on splanchnic blood flow have been described. Auzinger and coworkers 64 studied seven patients with catecholamine-refractory septic shock and subsequent infusion of terlipressin using gastric tonometry. During the 24-hour intervention period, terlipressin was administered as an intermittent bolus (1–3 mg). Although no changes occurred in lactate levels, the PCO₂ gap progressively increased over 72 hours.

Both vasopressin and terlipressin are potent vasoconstrictors and both are able to restore blood pressure in vasodilatory or septic shock. However, the effects on splanchnic blood flow are not yet fully elucidated. Clearly, adequacy of volume resuscitation is a major prerequisite for maintenance of microcirculatory blood flow. The currently available data suggest that both substances administered to compensate for endogenous vasopressin deficiency may be beneficial. Although the armamentarium for treatment of septic shock is enriched by such substances, it remains unclear whether administration during septic shock decreases morbidity or improves survival, and further research is warranted.

Modulation of the cytokine response by catecholamines might be a mechanism by which decreased morbidity and mortality are achieved with supranormal oxygen delivery in high-risk surgical patients 65. Phosphodiesterase III inhibitors have positive inotropic, vasodilating and anti-inflamatory properties, and they may avoid the development of tolerance to catecholamines as a result of β-receptor desensitization.

In a prospective, double-blind study 66, 44 patients with septic shock and conventional resuscitation were randomly assigned to receive dobutamine or enoximone to maximize left ventricular stroke work index. At 12 and 48 hours after baseline measurements, liver blood flow was assessed with hepatic venous catheterization, liver function was derived from appearance in plasma of MEGX, and release of tumour necrosis factor-α was determined to assess the severity of ischaemia/reperfusion injuries. There was a similar increase in cardiac index, systemic oxygen delivery and consumption, and liver blood flow in the two groups. Fractional splanchnic blood flow decreased slightly but significantly in dobutamine-treated patients, whereas it remained unchanged in enoximone-treated patients. In the latter group liver oxygen consumption and MEGX kinetics were significantly higher at 12 hours but not at 48 hours. The release of hepatic tumour necrosis factor-α after 12 hours of dobutamine treatment was twice as high (P < 0.05) as during enoximone treatment, suggesting a faster anti-inflammatory effect of enoximone. These interesting findings on hepatosplanchnic effects of phosphodiesterase III inhibitors were not confirmed by other studies, and further investigations are needed if these agents are to be recommended for routine clinical use.

Prostacyclin or its stable analogue iloprost are vasodilator substances with platelet aggregation inhibiting and cytoprotective properties. Administration of prostacyclin by the intravenous route was shown to increase oxygen delivery and consumption in septic patients 67 and to improve gastric pHi 68, as did aerosolized prostacyclin in patients with septic shock and pulmonary hypertension treated with epinephrine or norepinephrine 69. Finally, Lehmann and coworkers 70 reported restored plasma ICG clearance without harmful effect on systemic haemodynamics in patients with septic shock treated with iloprost.

More recently Kiefer and colleagues 71 reported the hepatosplanchnic effects of iloprost in 11 patients with septic shock requiring norepinephrine. Iloprost was incrementally infused to increase cardiac index by 15%, which significantly increased splanchnic blood flow in parallel, without a major fall in mean arterial pressure. Iloprost induced a decrease in endogenous glucose production rate without change in the hepatic clearance of the glucose precursors alanine, pyruvate and lactate. Similarly, the PCO₂ gap was not altered. The authors avoided mean arterial pressure drop by careful exclusion of hypovolaemia before inclusion, but still the increment in iloprost doses was limited by the decrease in arterial partial oxygen tension, which raises many questions in patients with acute respiratory distress syndrome. These interesting findings on hepatosplanchnic effects of such vasodilators need further investigation before these agents may be recommended for routine clinical use 72.

Opening the microcirculation using a vasodilator is an alternative approach for treatment of the jeopardized microcirculation in patients with sepsis or septic shock. Data reported by Sprock and coworkers 73 suggest that the use of intravenous nitroglycerin results in improved sublingual microvascular flow, as assessed by orthogonal polarization spectral imaging. However, one cannot assume that the sublingual microcirculation necessarily behaves like the whole splanchnic microcirculation does.

N-acetyl cysteine (NAC) administration was associated with a decrease in gastric pHi in septic patients 7475 and prevented the decrease in pHi in septic patients under hyperoxic stress 76. In a randomized, double-blind study conducted in septic shock patients, NAC given within the first 24 hours after admission to the ICU was shown to improve cardiac index and splanchnic blood flow and MEGX concentration, and to decrease gastric mucosal PCO₂ gap, whereas it did not influence fractional splanchnic blood flow 75. Nevertheless, these positive effects of NAC on the splanchnic circulation must be balanced against several negative studies. Indeed, NAC was reported to depress cardiac performance in septic patients 77, and it even worsened mortality rate when it was given more than 24 hours after hospital admission 78. Is NAC a 'double edged sword'? This question should be answered before its use in daily practice can be recommended.