Septic shock is a syndrome characterized by tissue hypoperfusion and hypotension secondary to an uncontrolled infection. It is a frequent cause of admission to the intensive care unit (ICU) and has an associated mortality around 40%. Around 50 % of septic shock patients exhibit early acute kidney injury and 30 to 40% will require renal replacement therapy.

After initial fluid resuscitation most of the patients with septic shock become hyperdynamic but still require norepinephrine (NE) to maintain a mean arterial pressure (MAP) above 65 mmHg. The optimal perfusion pressure may vary, specially in previously hypertensive patients as they may have a shift to the right in their kidney auto-regulatory curve. In a previous study in patients with chronic hypertension and septic shock, increasing MAP from 65 mmHg to 85 mmHg with NE was associated with improved renal function. However, the incidence of tachyarrhythmias increased, associated to the higher NE doses required, which has raised some concerns about the safety of this strategy. In this setting, the addition of vasopressin (AVP), a drug used as a vasopressor but with cathecholamine independent mechanisms, may allow to prevent this side effect by decreasing NE dose requirements. Low doses of AVP appear to be safe and when combined with NE in septic shock patients, it resulted in increased creatinine clearance and decreased use of renal replacement therapy, compared to NE alone. Theoretically, AVP can improve glomerular filtration rate. Therefore, the addition of AVP to NE in previously hypertensive septic shock patients should be a reasonable strategy to improve organ perfusion.

Furthermore, AVP could be an important step towards decatecholaminization in the management of septic shock patients. However, its effect on cardiac performance and stroke volume when targeting high MAP is unclear.

A) Theoretical foundations and state of the art A.1 Introduction Septic shock is a syndrome characterized by tissue hypoperfusion and hypotension secondary to an uncontrolled infection. It is a frequent cause of admission to the intensive care unit (ICU) and has an associated mortality around 40%(1). Several infectious diseases may lead to septic shock including different agents (bacteria, virus or fungi) and different sites (e.g. Pneumonia, abdominal infections, or urinary tract infections). Even Covid-19 may cause septic shock in a significant proportion of patients admitted to ICU(2).

Acute kidney injury (AKI) is a major complication associated to septic shock. Around 50 % of septic shock patients exhibit early AKI and 30 to 40% will require renal replacement therapy(3, 4), both factors associated with worse outcomes. Until now, there are no specific treatments to prevent AKI or renal failure. The current approach is based on optimizing systemic hemodynamics and promptly correcting hypoperfusion by the administration of fluids and vasopressors.

Tissue hypoperfusion is often present in septic shock patients due to several mechanisms including systemic vasodilatation, relative hypovolemia, myocardial depression, endothelial and microcirculatory dysfunction, and low arterial pressure with impaired global perfusion pressure. The hemodynamic treatment of septic shock is aimed at both, maintaining oxygen delivery above a critical threshold, while keeping a MAP at a level that allows adequate organ perfusion(5). Organ auto-regulation plays an important role to maintain organ blood flow over a range of perfusion pressures(6). There is evidence that that local auto-regulation may be impaired in septic shock(7). The optimal perfusion pressure may vary, especially in previously hypertensive patients as they may have a shift to the right in their kidney auto- regulatory curve(8). This concept was highlighted by a landmark study performed by Asfar and cols.(4), which compared the use of a higher target of MAP (80-85 mmHg) vs the usual target (65-70 mmHg) in septic shock patients treated with NE as the sole vasopressor. Although the study found no differences in the whole study group, among patients with chronic hypertension, those assigned to the higher MAP target had an improved renal function. A Consensus on circulatory shock and hemodynamic monitoring of the European Society of Intensive Care medicine (ESICM 2014) recommended to titrate NE to higher MAP goals in previously hypertensive patients(9). However, in the study of Asfar the incidence of atrial fibrillation increased in the group assigned to the high MAP target, associated to the higher NE doses required, which has raised some concerns about the safety of this strategy(4). In this context, the addition of exogenous vasopressin (AVP), a drug used as a vasopressor but with cathecholamine independent mechanisms, may prevent this side effect by decreasing NE dose requirements.

Low doses of AVP in septic shock appear to be safe and when combined with NE to maintain a conventional target of MAP, it resulted in increased creatinine clearance and decreased use of renal replacement therapy, compared to NE alone (10). Theoretically, AVP can improve glomerular filtration rate(11). Therefore, the addition of AVP to NE to target a higher MAP goal in previously hypertensive septic shock patients may be a reasonable strategy to preserve renal function while avoiding the side effects of high doses of cathecolamines. However, in contrast to NE, AVP increases systemic vascular resistance (afterload) without a parallel inotropic effect. Myocardial dysfunction is highly prevalent in septic shock and it may contribute to persistent hypoperfusion and to worse outcomes(12, 13). Therefore, there is a reasonable concern regarding the impact of AVP on cardiac performance and stroke volume when targeting higher blood pressures.

A.2 Renal dysfunction in septic shock Around 80% of septic shock patients develop AKI throughout their evolution and 30 to 40% require renal replacement therapy (4, 14). According to severity, AKI is usually graded in 3 stages(15). A large epidemiologic study showed that critically ill patients meeting stage 3 AKI criteria have an associated mortality of 51.1%, and in those patients who required renal replacement therapy (RRT) it was 55.3%(16). A large clinical trial of septic shock revealed that 50.4% of patients already had stage 2-3 AKI at ICU admission. Among patients without AKI at enrollment, 37.8% developed AKI during their evolution and the requirement of RRT was around 6%. Moreover, 60-day mortality was three to five times higher in those who developed AKI, independent of whether AKI was present at admission or developed later (3). In addition, AKI adversely impacts short- and long-term clinical outcomes, and healthcare costs (17, 18).

No specific resuscitation strategy appears to influence AKI progress in patients with septic shock(3). Regardless of the improvement in systemic hemodynamics, renal dysfunction is unresponsive to aggressive resuscitation with fluids and NE in experimental settings(7). The status of renal blood flow is unclear since it has been reported as increased, decreased, or unchanged in human sepsis. Data on renal blood flow from animals is heterogenous and does not provide clear conclusions. Nevertheless, cardiac output is the most important independent predictor of renal blood flow in sepsis(19). In some patients, an increase in RBF is associated with a decrease in glomerular filtration rate (GFR). This could be potentially explained by efferent arteriolar vasodilatation(20), persistent intra renal shunting (microcirculation abnormalities) and/or failure of autoregulation (7). In a sheep model of sepsis-induced kidney injury, manifest reductions of renal vascular resistance were found, suggesting auto regulation failure. The increase in MAP to 70 mmHg and normalization of systemic hemodynamics by means of fluids and NE failed to improve renal blood flow (RBF) and cortical microcirculation(7). However, there was no attempt to further increase blood pressures.

A.3 Physiologic and clinical evidence for higher blood pressure targets in Septic Shock The physiologic impact of targeting higher MAP levels on organ perfusion in septic shock has shown conflicting results. Thooft et al.(21) found that increasing MAP with NE during septic shock can increase cardiac output and improve sublingual microvascular flow (as marker of tissue perfusion) in stable resuscitated patients. Dubin et al.(22) increased MAP with NE in a cohort of septic shock patients and observed no improvement in mean sublingual microcirculatory indexes. However, there were considerable variations in the inter-individual responses, depending of the basal condition of the microcirculation. Likewise, Bourgoin et al.(23) showed that increasing MAP from 65 to 85 mm Hg with NE, during a four- hour study period in septic shock patients, resulted in an increase in cardiac output, systemic vascular resistance, and in both, left and right ventricular stroke work indexes, but with no significant changes in oxygen consumption, lactate or urine output.

The first large randomized trial that assessed the impact of targeting higher blood pressures in septic shock is the study of Asfar and cols.(4), mentioned above. According to their hypothesis, which stated that higher perfusion pressures could be associated to improved outcomes, especially in previously hypertensive patients, randomization was stratified according to whether patients had previous chronic hypertension or not. In the whole group of (n=776) and in the subgroup of patients without chronic hypertension (n=336), no differences were observed in renal function. However, among the 340 patients with chronic hypertension, those assigned to the high MAP target (80-85 mmHg) exhibited a decreased in both the incidence of AKI, defined as a doubling of the blood creatinine level, and the rate of renal-replacement therapy (RRT)(4). Nonetheless, the incidence of atrial fibrillation was significantly higher in the patients assigned to the higher-target group, probably because they required significantly higher doses of NE to achieve a MAP of 80-85 mmHg(4).

Some recent studies have proposed to apply the high MAP target as a short test for septic shock patients with evidence of persistent hypoperfusion, with or without chronic hypertension. The ANDROMEDA-SHOCK trial recommended a 2-hour "vasopressor test" directed to improve tissue perfusion(24), and maintained the high blood pressure only if evidence of a positive result. Because of the design of the study it is not possible to conclude whether this specific intervention improved outcomes. An ongoing study, TARTARE- 2S(25), proposes a similar vasopressor test for 2 hours in patients with septic shock and oliguria, maintaining the higher MAP target in case of positive response.

Until now, all the studies targeting higher blood pressures in septic shock have used NE as the sole vasopressor. The doses required to increase MAP from 65 to 85 mmHg are highly variable but may be as high as two to three - fold.

A.4 Physiologic effects of Norepinephrine in patients with septic shock Norepinephrine (NE) is the vasopressor of choice in septic shock(26). It has broad effects on different determinants of cardiac and renal function. It is an alpha-adrenergic agonist that increases vascular tone, which may impact both preload and afterload, although the former effect seems to predominate in septic shock patients. Monnet and cols.(27) studied the impact of NE in a group of septic shock patients with evidence of preload dependence and observed that NE increased left ventricular diastolic volumes and cardiac output, while decreasing preload dependence. Persichini and cols.(28) demonstrated that this increase in venous return is due to an increase in mean systemic filling pressures (Pmsf) and the driving pressure for venous return (=Pmsf - right atrial pressure). By inducing venoconstriction in the capacitance vessels, NE recruits volume from the unstressed to the stressed compartment increasing Pmsf.

A second potentially favorable effect of NE is the increase in the contractile force of the myocardium, mainly due to activation of beta 1 adrenoceptors, although an alpha-1 adrenoceptor-mediated inotropic response can be also elicited by NE(29). Hamzaoui et al.(30) showed that NE increased left ventricular ejection fraction and stroke volume in 38 septic shock patients. This positive inotropic effect was also observed in patients with a left ventricular ejection fraction (LVEF) < 45%. The combination of the rise in arterial pressure and contractility may improve ventriculo-arterial coupling(31). Nonetheless, heart rate is not modified significantly by NE. Therefore, increased cardiac output in response to NE in septic shock seems to be the result of the combined effects of NE on preload and contractility.

The increase in afterload elicited by NE has shown to be potentially deleterious for cardiac performance in postoperative cardiac patients(32). However, in septic shock the increase in afterload does not appear to worsen cardiac function. In our preliminary unpublished data (see work in progress and Figure 1 and Table 1 of supplementary material) in 12 septic shock patients studied with a comprehensive assessment by echocardiography, the investigators found no evidence of cardiac dysfunction when increasing NE to target a higher blood pressure.

Concerning renal function, although old experimental physiologic studies indicated that NE could decrease renal blood flow secondary to increased renal vascular resistance (33), more recent studies showed that NE consistently increases renal blood flow, glomerular filtration rate and urine output when used in septic shock patients (34, 35). Renal vascular resistance is typically decreased in septic shock(7, 36). NE may restore this resistance while increasing renal perfusion pressure as a result of increased arterial pressure.

A.5 Physiologic effects of vasopressin (AVP) in patients with septic shock Although NE is still recommended as the first-line vasopressor in septic shock(26), during the last decade there has been growing interest in the use of AVP as an adjunctive agent(37), especially after the release of several studies reporting relative endogenous vasopressin deficiency in patients with septic shock, and that such deficiency may contribute to the diminished vascular tone observed in septic shock(38). Vasopressin is the mediator of a remarkable regulatory system for the conservation of water. The hormone is released by the posterior pituitary gland whenever water deprivation causes an increased plasma osmolality or whenever the cardiovascular system is challenged by hypovolemia and/or hypotension. Vasopressin effects are mainly due to its interactions with the 3 types of receptors, V1a (vasoconstriction), V1b (ACTH release), V2 (anti-diuretic effects). AVP is a potent vasoconstrictor primary mediated by the V1a receptor found in the vascular smooth muscle(39). Since splanchnic vasodilation is a main pathophysiologic characteristic of septic shock, vasoconstriction of the splanchnic vasculature induced by AVP may increase stressed volume and improve venous return. In addition, stimulation of arterial smooth muscle may increase systemic vascular resistance with further increase of MAP.

The hemodynamic effects of low doses of AVP (0.05 ± 0.02 U/min) in septic shock have been analyzed in a meta-analysis, which included more than 900 patients with septic shock from 9 randomized controlled trials. Compared to control patients treated with NE alone, those treated with AVP (+ NE) exhibited a 12% mean decrease in heart rate, a 14% increase in stroke volume, and no change in cardiac output. Regarding the effect of AVP on NE requirements, use of AVP was associated to a decrease in NE dose of 1.6 times. None of the studies included in this meta-analysis applied a high target of MAP(40). One randomized controlled study in 50 septic shock patients compared two doses of AVP: 0.03 U/min vs 0.06 U/min and observed that the 0.06 dose was associated to a larger decrease in NE requirements without increased adverse effects(41). A more recent large randomized controlled trial (VANISH) confirmed the hemodynamic effects described in the meta-analysis(10). The subtle decrease in heart rate observed in most studies is not physiologically surprising. AVP is known to act on the area postrema to enhance the baroreflex activity(39). Vasoconstriction induced by AVP is mediated by activation of V1a-receptor in the vascular smooth muscle. This pathway differs with the mechanism of action of catecholamines and may explain why AVP complements NE in septic shock(42). Some studies have measured the effects of AVP compared to NE alone on perfusion variables (lactate levels and gastric CO2) and no changes have been observed(11, 43). Regarding cardiac complications, no significant differences in acute myocardial infarction or cardiac arrest have been observed in the larger trials(10, 44). In this context, low doses of AVP appear to be safe in septic shock and not associated with hemodynamic deterioration. However, in septic shock patients the effects of AVP on cardiac function, assessed with echocardiography, has not been evaluated.

Regarding the effects of AVP on renal function in septic shock, several clinical studies have suggested that AVP might be superior to NE in maintaining glomerular filtration rate and improving creatinine clearance (10, 11, 43, 44) Patel and cols(11). performed a randomized controlled double-blinded study in 24 septic shock patients which showed that a 4-hour infusion of AVP (0.06 U/min) over basal NE increased creatinine clearance in 75% and doubled urine output, while no changes were observed in the group treated with NE alone. In a similar study in 23 septic shock patients, Lauzier et al(43). observed that infusion of AVP (0.04 - 0.2 U/min) doubled creatinine clearance after 24 hours, compared to no change in the group treated with NE alone. The VASST trial, a RCT comparing NE alone to NE plus AVP at 0.03 U/min in 778 septic shock patients, showed no difference in the primary outcome of mortality (44). However, in a post hoc analysis it was found that in the 106 patients classified as to be in risk of AKI at baseline, treatment with AVP was associated to lower rate of progression to renal failure and renal replacement therapy(45). In the VANISH trial, 409 patients with septic shock were randomized to receive AVP (0.06 U/min) or NE alone. Although there was no difference in the primary outcome (days free of renal failure), the group treated with AVP required less renal replacement therapy (25.4% vs 35.3%) (10). It has been proposed that the positive effect of AVP on renal function would be due to selective vasoconstriction of the glomerular efferent arteriole mediated by V1 receptor activation(46), and by vasodilation of afferent arterioles mediated by V2 receptor activation(47).

A.6 Our proposal A recent worldwide survey among 839 intensivists demonstrated that almost 70% of respondents added a second vasopressor over NE because of three reasons: to limit/reduce the side-effects of the first vasopressor, to use a second drug with an independent mechanism of action, or to achieve a synergistic effect between both drugs (48). Current guidelines (2016) recommend, although admittedly with a low level of evidence, to add AVP to NE to achieve MAP targets or to decrease NE doses in septic shock (26). Despite this recommendation, there is still not a clear indication for its use.

In the particular setting of septic shock patients with chronic hypertension, which represent around 30 to 40% of all septic shock patients, the kidney auto-regulation curve can be shifted significantly to the right (8). This selected group of patients in which higher targets of MAP have been proposed, which is associated to a marked increase in the required doses of NE, may benefit of adding AVP to the vasopressor approach. However, there are no reports of AVP use in this specific indication.

The investigators propose a randomized controlled trial of AVP vs placebo to target a higher MAP in septic shock patients with chronic hypertension, with a comprehensive clinical and physiologic assessment of renal and cardiac function.

Inclusion Criteria:

• Septic shock diagnosed at ICU admission according to the Sepsis-3
• Mechanical ventilation in place
• Past medical history of chronic hypertension
• Fluid unresponsive status
• Stable norepinephrine dose ≥ 0.1 mcg/kg/min
• Persistent tissular hypoperfusion after initial resuscitation

Exclusion Criteria:

• Age < 18 years
• > 24 h since septic shock diagnosis
• Moderate or severe mitral/aortic disease
• Anticipated surgery during the study period
• Abdominal hypertension grade III
• Pregnancy
• Do-not-resuscitate status

The statisticians and researcher responsible to perform the measurement will be blinded to the group allocation. Nonetheless, the research nurse responsible for event assignment will not be blinded to the group allocation.