Intravenous angiotensin II for the treatment of high-output shock (ATHOS trial): a pilot study

Critically ill patients with shock requiring vasopressors are at a high risk of death. Distributive shock is the most common form, and is often caused by sepsis [1]. When shock is treated with vasopressors, two main classes of vasopressors are in the intensivists’ armamentarium: catecholamines and vasopressin-type peptides [1]. Currently, no specific type of vasopressor (for example, norepinephrine, vasopressin, dopamine) compared to another vasopressor has been shown to improve outcome [2]. All vasopressors have limitations and potential side effects. Patients treated with catecholamines for shock often develop tachyphylaxis, thereby limiting the utility of these agents, and high doses of catecholamines can cause direct cardiotoxicity [3]. The toxic potential of catecholamines has been recently demonstrated in a randomized clinical trial of septic shock patients treated with norepinephrine [4]. In this study, beta blockade with esmolol was shown to improve survival in these patients by decreasing the heart rate. Thus, vasopressors that are not inotropes or chronotropes may be useful in patients with shock. One such vasopressor is vasopressin, which is most commonly used as an adjuvant with catecholamines. Vasopressin has been shown to improve outcomes in patients with less severe septic shock, but has toxicity (that is, cardiac and mesenteric ischemia) at high doses and interacts with hydrocortisone [5]. In the subset of critically ill patients in whom mean arterial pressure (MAP) cannot be maintained with vasopressors, distributive shock is uniformly fatal. The addition of a rescue vasopressor in this setting could be useful.

Angiotensin II (ATII) is a naturally occurring hormone with endocrine, autocrine, paracrine, and intracrine hormonal effects. It is a potent direct vasoconstrictor, constricting both arteries and veins and increasing blood pressure [6]. It has a half-life in circulation of approximately 30 seconds, but while in tissue, its half-life may be as long as 15 to 30 minutes. Importantly, ATII increases secretion of antidiuretic hormone (ADH) and adrenocorticotropin hormone (ACTH), and may potentiate sympathetic effects by direct action on postganglionic sympathetic fibers. It also acts on the adrenal cortex, causing it to release aldosterone [6],[7]. We hypothesized that ATII might serve a role as a useful vasopressor in the treatment of shock, but the appropriate dose of intravenous ATII to increase blood pressure is unknown. The dose of intravenous ATII previously described has varied across studies, but ranges from 0.4 ng/kg/min to as much as 40 ng/kg/min. The highest doses were reported in the cases of profound hypotension due to angiotensin-converting-enzyme inhibitor (ACEi) overdose [8],[9]. We set out to determine the appropriate dose of ATII in the treatment of high output shock.

The study was conducted at the George Washington University Hospital Intensive Care Unit, Washington DC, USA. The trial was registered on clinicatrials.gov (NCT01393782) and the study protocol was approved by the Food and Drug Administration (IND# BB-IND-11592). The protocol was approved by the George Washington University Institutional Review Board. Written informed consent was obtained from each participating patient or appropriate surrogate prior to enrollment. The investigators performed all experimental procedures and the study coordinators recorded the data.

The flow of patients into the study is reported in Figure 2: 20 patients underwent randomization and all 20 patients were enrolled in and completed the study (Figure 1). Baseline characteristics of the two groups are shown in Table 1. The mean age for all study subjects was 62.9 ± 15.8 years. Of the patients, 75% were male, 45% were Caucasian and 40% were African American. Baseline SOFA and acute physiology and chronic health evaluation II (APACHE II) scores were 15.9 ± 3.0 and 30.6 ± 8.9, respectively. Of the 20 patients 19 were receiving concomitant vasopressin at a dose of 0.02 to 0.08 u/min. Vasopressin doses were not adjusted during the study period.

ATII resulted in a reduction in norepinephrine dosing in all patients (Figure 3). The mean hour-1 norepinephrine dose for the placebo cohort was 27.6 ± 29.3 mcg/min versus 7.4 ± 12.4 mcg/min for the ATII cohort (P = 0.06). Hour 2 norepinephrine dosing for the placebo cohort was 28.6 ± 30.2 mcg/min versus 7.3 ± 11.9 mcg/min in the ATII cohort (P = 0.06). Throughout the study period, the mean ATII dose was reduced from 20 ng/kg/min at hour 0 to 5 ng/kg/min at hour 6 before being titrated off by hour 7 (one hour post-infusion). Despite this down-titration of ATII, norepinephrine doses remained substantially lower in the ATII cohort than the placebo cohort, though the effect approached statistical significance only at hours 1 and 2. Upon cessation of the ATII infusion, mean norepinephrine rebounded concomitantly.

Using a general estimating equation model with time defined as a continuous variable, in order to obtain a global test of interaction effect, the main effect of treatment (study drug versus placebo) was not significant (P = 0.13), nor was the effect of time (P =0.30), nor was the treatment multiplied by time interaction (P = 0.76). When time was defined as a class variable with hour-1 defined as the reference group, in order to examine specific time points, the drug effect (P = 0.14) and time effect (P = 0.18 at time 0, P = 0.51 at time 1) both remained non-significant. The product of drug multiplied by time interaction showed a trend level of significance at 1 hour and 2 hours (P = 0.06).

Adverse events most commonly experienced by all patients were metabolic disorders with alkalosis occurring in four patients in the ATII group and no patients in the placebo group (P = 0.09). The most common adverse event thought to be attributable to ATII was hypertension, which occurred in 20% of patients receiving ATII (P = 0.58). In both of these patients, the study drug infusion was stopped, per protocol, in order to achieve MAP goals. Table 2 lists adverse events.

Urine output, cardiac output, central venous pressure, and MAP are shown in Table 3. The 30-day mortality for the two groups was similar for the ATII cohort and the placebo cohort (50% versus 60%, P = 1.00).

We report the findings of the first prospective randomized placebo-controlled trial of ATII in the treatment of high-output shock. Our efforts were intended as a proof-of-concept and dose-finding study, as well as an attempt to generate hypotheses in advance of larger future studies. To this end, we have shown that ATII can be an effective pressor agent at a dose range of 5 to 40 ng/kg/min. More specifically, we believe that a starting dose of 2 to 10 ng/kg/min may be an appropriate starting dose in the treatment of high-output shock when used in conjunction with standard-of-care vasopressors.

ATII has been used previously for the treatment of hypotension in a handful of cases. Newby et al. describe the successful treatment of a patient with an ACEi overdose and profound shock, using an ATII infusion [11]. Multiple case reports have shown potential utility for ATII in the treatment of septic shock with catecholamine infusions [12-16].

While all patients in the present study had a response to the ATII infusion, we observed significant heterogeneity. Of the ten patients who received ATII, two had a modest response, while two were exquisitely sensitive to ATII, which was an unexpected finding. In the two highly sensitive patients, the norepinephrine infusion was titrated off per protocol, and the ATII dose was at its lowest allowable dose of 5 ng/kg/min and the patients remained hypertensive with MAP of >90 mm Hg, despite norepinephrine being titrated off. As hypertension is not part of our standard of care, the investigators halted the infusion, and the ATII was weaned off. In both cases the need for norepinephrine was rapidly re-established. Pre-clinical studies demonstrate that animals that become septic after pre-treatment with enalapril are resistant to norepinephrine as a vasopressor [17]. Based on these findings, we hypothesized that the two patients who were very sensitive to ATII may have been receiving an angiotensin-converting enzyme inhibitor prior to developing septic shock, thus, explaining the exquisite sensitivity to ATII infusion. However, a detailed chart review revealed incomplete information, and we were unable to document pre-morbid exposure to angiotensin-converting enzyme inhibitor. Therefore, the ATII sensitivity that we observed could be due to other mechanisms that have not been elucidated.

We observed that ATII may have synergy with other vasopressors (that is, catecholamines and vasopressin), but may also have another important indication in critically ill patients. Previous work in pre-clinical studies suggests that septic animals suffer acute kidney injury in part due to intra-glomerular hypotension induced by efferent arteriole vasodilation. In these models, intravenous infusion with ATII restores creatinine clearance and urine output [18]. While the present study was underpowered to elucidate any effect on urine output, we expect further large-scale studies will clarify the effect of ATII on kidney function in high-output shock. In addition, ATII is a potent vasopressor without inotropic or chronotropic properties [19]. Recent randomized controlled trials in patients with septic shock treated with norepinephrine suggest that less chronotropy may be desirable and may lead to a survival benefit [4]. Based on these findings, for patients who require norepinephrine and are tachycardic, ATII may be particularly useful. We hypothesize that for patients with severe hypotension, lower doses of multiple vasopressors with differing mechanisms of action may be more efficacious and less toxic than high doses of one type of vasopressor (that is, catecholamines).

The study had multiple strengths. First, the study was a randomized double-blind controlled trial with an appropriate placebo-control arm. Second, it was of pragmatic design, as it was the intent of the investigators to enroll patients receiving standard-of-care treatment for high-output shock. As such, all patients had received a priori appropriate monitoring and therapeutic interventions (including central venous lines, bladder catheters, arterial lines, and cardiac output monitoring devices). There was no additional need for any specialized equipment of procedures prior to enrollment in the study. Third, all enrolled patients had a documented need for high-dose vasopressor therapy despite volume therapy, as evidenced by the cardiac index entry criteria. This, we believe, is in keeping with the current practice of addressing volume responsiveness in a hypotensive patient prior to initiation of vasopressor therapy. Finally, as part of our study protocol, we employed the use of a data safety monitor, who had the ability to unblind data and evaluate for adverse events as well as halt the study, neither of which occurred.

The study had several limitations. First, the study had a modest sample size. As such, we were unable to make conclusions about the effect of ATII on urine output due to the high incidence of oliguria and renal replacement therapy in both the ATII group and the placebo group. Moreover, this study was not powered adequately to discern a difference in mortality between the ATII and placebo groups. Second, there were some imbalances between our placebo arm and our drug arm, the former of which were younger, but sicker (according to both SOFA and APACHE II scores). It is possible that differences in these two populations influenced the effectiveness of ATII. Finally, inclusion criteria for enrollment in the study were such that the resulting study population was critically ill, with an expected mortality in excess of 50%. Indeed, our requirement of a cardiovascular SOFA score of 4 (indicating a norepinephrine dose of 0.1 mcg/kg/min) would, on average, equate to a minimum norepinephrine dose of 9.3 mcg/min, which based on our study drug titration protocol (Figure 1), would leave little room for titration. In order to allow for the possibility of a meaningful signal upon initiation of ATII, we preferentially considered patients with a substantially higher starting dose of norepinephrine (20 to 30 mcg/min, as evidenced in Figure 3). Moreover, we preferentially considered patients with an upward-trending norepinephrine requirement, signifying refractoriness to therapy. Based on these facts, our results may not be generalizable to a less sick population. However, we foresee a use for ATII in a significant critically ill population, for whom multiple vasopressors are required.

The initiation of an ATII infusion in patients receiving norepinephrine for septic shock resulted in a marked decrease in norepinephrine doses. ATII may be effective as a novel pressor agent in the treatment of high-output shock. Initial dosing ranges are most likely between 2 and 10 ng/kg/min. In our pilot study, the drug appears to be well-tolerated. Further randomized placebo-controlled trials to more fully elucidate the role of ATII as a vasopressor in the treatment of shock are warranted.

Angiotensin II improved blood pressure in patients with high-output shock and multiple vasopressors.

Angiotensin II may have a role as a rescue vasopressor.

The response to angiotensin II in a modest-sized study were highly variable.

The precise role of angiotensin II in the treatment of hypotension and shock needs further study.

Angiotensin II was well-tolerated in this pilot study.

LSC conceived the idea, obtained regulatory approval, designed the study, wrote the protocol, led the conduct of the study, helped write the manuscript, produced tables and figures, and gave final approval of the version to be published. LWB assisted with regulatory approval, made substantial contributions to the acquisition of data and data interpretation, and was involved in drafting the manuscript and revising it critically for important intellectual content. EBM helped conduct the study, made substantial contributions to the acquisition of data and data interpretation, and was involved in drafting the manuscript and revising it critically for important intellectual content. DD helped conduct the study, and was involved in drafting the manuscript and revising it critically for important intellectual content. JH performed safety monitoring, and was involved in drafting the manuscript and revising it critically for important intellectual content. ZA helped conduct the study, and was involved in drafting the manuscript and revising it critically for important intellectual content. MGS helped conduct the study, and was involved in drafting the manuscript and revising it critically for important intellectual content. All authors have read and approved the final manuscript.