A Multicenter Observational Cohort Study of Angiotensin II in Shock

Abstract

Introduction:

Angiotensin II (Ang-2) is a non-catecholamine vasopressor that targets the renin-angiotensin-aldosterone system by agonism of the angiotensin type 1 receptor. Its utility as a vasopressor and a catecholamine-sparing agent was demonstrated in the pivotal ATHOS-3 trial, and numerous post-hoc analyses have shown reduced mortality in certain subsets of the population.

Methods:

Consecutive adult patients at 5 centers who received Ang-2 from 2017–2020 were included in this multicenter, retrospective observational cohort study. Patient demographics, hemodynamics, and adverse events were collected. The primary outcomes of the study were the mean difference in MAP and norepinephrine (NEpi)-equivalent dose at hours 0 and 3 following initiation of Ang-2 therapy.

Results:

One hundred and sixty-two patients were included in this study. The primary outcomes of an increase in MAP (mean difference 9.3 mmHg, 95% CI 6.4–12.1, p < 0.001) and a reduction in NEpi equivalent dose (mean difference 0.16 μg/kg/min, 95% CI 0.10–0.22, p < 0.001) between hours 0 and 3 were statistically significant. The median time to reach a MAP ≥65 was 16 minutes (IQR 5–60 min). After stratifying patients by the NED dose and number of vasopressors administered prior to the initiation of Ang-2, those with a NED dose < 0.2 μg/kg/min, NED dose < 0.3 μg/kg/min, or those on ≤ 3 vasopressors had a significantly greater reduction in NED by hour 3 than those patients above these thresholds.

Conclusion:

Ang-2 is an effective vasopressor and reduces catecholamine dose significantly. Its effect is rapid, with target MAP obtained within 30 minutes in most patients. Given the critical importance of adequate blood pressure to organ perfusion, Ang-2 should be considered when target MAP cannot be achieved with conventional vasopressors. Ang-2 should be utilized early in the course of shock, before the NED dose exceeds 0.2–0.3 μg/kg/min and before the initiation of the fourth-line vasopressor.

Introduction

Angiotensin II (Ang-2) is a non-catecholamine vasopressor approved by the Food and Drug Administration (FDA) in 2017 and by the European Medicines Agency (EMA) in 2019 for the treatment of patients with distributive shock.¹ Initially studied in the Angiotensin for the Treatment of High-Output Shock (ATHOS) and ATHOS-3 trials, Ang-2 was found to be effective at raising mean arterial blood pressure (MAP) and decreasing norepinephrine (NEpi)-equivalent dose requirements.^2,3

Ang-2 is a naturally occurring hormone in the renin-angiotensin-aldosterone system (RAAS) and is an agonist of angiotensin type 1 (AT₁) receptors on vascular smooth muscle cells.⁴ Agonism of AT₁ receptors stimulates calcium/calmodulin-dependent phosphorylation of myosin light chain kinase, resulting in smooth muscle contraction and vasoconstriction.⁴ Additionally, Ang-2 stimulates synthesis of cortisol, exocytosis of renin, adrenergic stimulation of cardiomyocytes, and aldosterone secretion, which results in increased sodium and water reabsorption. Collectively, these actions result in increased circulatory volume and MAP.⁵

Numerous studies have examined the use of bovine Ang-2 since the 1930s.^6–8 Since the ATHOS and ATHOS-3 trials, several subgroup analyses have been performed of ATHOS-3, which demonstrated mortality benefit in particular patient populations.^9–12 With the exception of case reports, small case series, and 1 multicenter study, studies occurring after ATHOS-3 that have not been derived from that study’s dataset are sparse.^13–17 The purpose of our multicenter study was to assess the effectiveness of Ang-2 for the treatment of shock. Specifically, we sought to evaluate whether Ang-2 effectively increases MAP and decreases NEpi-equivalent dose. Because Ang-2 agonizes AT₁ receptors in the RAAS, a distinct auto-regulatory system from the catecholamine system, we hypothesized that it would significantly increase MAP and decrease background NEpi-equivalent dose. In addition, we sought to characterize the timing of these effects, the types of shock in which it was being utilized, as well as the incidence of adverse events such as venous thromboembolic events (VTEs).

Methods

Patients

This multicenter, retrospective observational study was reviewed and approved by a central Institutional Review Board (IRB) at the University of Maryland. The manuscript was referenced against the STROBE checklist for cohort studies. Study sites consisted of 2 academic tertiary care hospitals, 2 private tertiary care hospitals, and one community hospital in the United States. In total, the aggregate capacity between the 5 sites totaled 2,792 hospital beds and 399 critical care beds across the medical, surgical, and cardiovascular intensive care units (ICUs). Consecutive adult patients admitted to each hospital between 2017 and 2020 who were administered Ang-2 for any form of shock were identified and included in the Angiotensin II Research to Evaluate the Multi Institutional Use in Shock (ARTEMIS) study. Patients presenting with all shock states were included in the analysis, including those receiving off-label Ang-2 for the treatment of non-distributive shock. Patients were excluded from our study if they were <18 years of age. Patient demographics, as well as data on hemodynamics, vasopressor dose, and outcomes were collected for all patients. Other characteristics collected included past medical history, exposure to an angiotensin converting enzyme inhibitor (ACE-I) or angiotensin receptor blocker (ARB), the etiology of shock, the types of infection, use of other rescue medications, in-hospital mortality, and the use of pharmacologic deep vein thrombosis (DVT) prophylaxis.

Study Outcomes

The study’s primary outcome was the mean difference between MAP and NEpi-equivalent dosing at the time of Ang-2 administration (hour 0) and 3 hours post Ang-2 initiation. NEpi-equivalent doses were calculated based on previous studies and are found in Table S1 of the Supplementary Appendix.³ Secondary outcomes were specified a priori and included the time to reach a MAP ≥65 mmHg or a MAP increase of 10 mmHg. The target MAP of 65 mmHg was utilized based on the Surviving Sepsis Campaign recommendations for septic shock management, while MAP increase by 10 mmHg was chosen to evaluate Ang-2’s efficacy as a vasopressor.¹⁸ In addition, the total duration of Ang-2 therapy, and the rate of MAP goal attainment within 15, 30, and 60 minutes of Ang-2 initiation were studied. Differences in hemodynamic parameters between those that responded and did not respond to Ang-2 were evaluated. Furthermore, the relative reduction in vasopressor dosage when stratifying patients by the NED dose and number of vasopressors prior to the initiation of Ang-2 were evaluated as exploratory outcomes.

Data related to vasopressor utilization were also evaluated, including time to vasopressor discontinuation after Ang-2 administration and order of vasopressor discontinuation. Renal outcomes such as the incidence of acute kidney injury (AKI) and utilization of renal replacement therapy (RRT) prior to Ang-2 administration, as well as the total days on RRT and the proportion liberated from RRT during hospitalization were studied. Acute kidney injury was defined according to the Acute Kidney Injury Network (AKIN) criteria, and any patient with an increase in serum creatinine by ≥0.3 mg/dL, increase in creatinine ≥ 1.5 times baseline, or urine output < 0.5 mg/kg/hr for 6 hours was classified as having AKI. Safety outcomes were also examined, and the incidence of VTE within 28 days of Ang-2 administration was calculated and categorized by severity of the event. This outcome was identified by detailed review of the electronic health record and confirmed by the principal investigator at each institution. The National Cancer Institute criteria for adverse events was used to define clinical severity in order to provide standardization across investigators collecting data.¹⁹

Statistical Analysis

Data analysis was performed using SPSS Statistics 25 (IBM Corporation, Somers, NY) and R 4.0.2 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables with a non-parametric distribution are reported as the median value and interquartile range (IQR). Categorical variables are reported as number and percentage. In order to highlight differences between hemodynamics and vasopressor dose at hours 0 and 3, the mean differences in hemodynamic parameters for continuous variables were reported and compared using the Wilcoxon signed-rank test. For all statistical inferences, a p-value <0.05 was considered statistically significant, all tests were 2-tailed, and p-values were adjusted for multiple comparisons utilizing a Bonferroni correction.

Given the retrospective nature of our study and unknown availability of data prior to analysis, an a priori sample size calculation was not performed. However, a post-hoc power calculation indicated that with an observed mean MAP of 62 mmHg prior to Ang-2 administration, our sample was large enough to detect an approximately 4.1 mmHg difference in MAP with an alpha of 0.05 and 80% power. In addition, with an observed mean NED dose of 0.63 μg/kg/min prior to Ang-2 administration, our sample was large enough to detect an approximately 0.16 μg/kg/min difference in NED with an alpha of 0.05 and 80% power.

Results

One hundred and sixty-two patients met study inclusion criteria. Median age was 63 (IQR 51–71) and 58.6% of patients were male (Table 1). Hypertension, congestive heart failure, and diabetes mellitus were the most common co-morbid conditions, and occurred in 68.4%, 50.0%, and 46.0% of patients, respectively. Exposure to an ACE-I or ARB prior to admission occurred in 21.7% and 13.2% of patients, respectively. The most common initial indication for vasopressor use was septic shock (72.2%), followed by cardiogenic shock (21.0%), non-septic distributive shock (14.2%), and hypovolemic shock (3.1%) (Table 1). In those patients with sepsis as the presumed etiology of shock, the most common organisms were gram positive (58.1%), followed by gram negative (10.3%), and fungal (8.5%). There was no growth in 17.1% of cultures, and culture data was unknown in 6.0% of cases. The location of infection occurred most frequently in the blood (35.0%), followed by sputum (23.9%) and urine (12.8%), and was unknown in 28.2% of cases.

Table 1.

Patient Characteristics.

Variable	(n = 162)
Demographics
Age, years	63 (51–71)
Male	95 (58.6)
Weight, kg	93 (76–108)
APACHE IV score	90 (71–121)
Past Medical History
HTN	68.4%
DM	46.0%
CAD	32.4%
CHF	50.0%
Renal Disease	45.3%
Liver Disease	25.2%
Home ACE-I	21.7%
Home ARB	13.2%
Etiology of Shock*
Distributive (septic)	72.2%
Distributive (non-septic)	14.2%
Hypovolemic	3.1%
Cardiogenic	21.0%
Obstructive	2.5%
Unknown	4.9%
Type of Infection
Gram negative	10.3%
Gram positive	58.1%
Fungal	8.5%
No growth	17.1%
Unknown	6.0%
Location of Infection
Blood	35.0%
Urine	12.8%
Sputum	23.9%
Unknown	28.2%
Ordering Location
MICU	67.3%
CVICU	17.9%
SICU	14.8%
Number of Vasopressors Prior to Ang-2
Median (IQR)	3 (2–3)
Distribution
0	2.5%
1	3.7%
2	25.3%
3	45.7%
≥ 4	22.8%
Baseline NEpi-equivalent dose
Median (IQR)	0.55 (0.38–0.82)
Distribution
<0.2	13.0%
≥0.2 to <0.5	30.9%
≥0.5	56.1%
Use of other rescue medications
Hydrocortisone	69.8%
Fludrocortisone	4.3%
Methylene Blue	5.8%
Hydroxocobalamin	7.9%
Ascorbic Acid	49.6%
Outcomes
ICU LOS	7 (2–17)
Hospital LOS	10 (4–26)
ICU mortality	78 (48.1)
In-hospital mortality	98 (60.5)

Continuous variables are represented as median (interquartile range).

^*Categories within etiology of shock exceed 100% due to mixed etiology in some patients.

Abbreviations: kg = kilograms, ACE-I = angiotensin converting enzyme inhibitor, ARB = angiotensin receptor blocker, ICU = intensive care unit, CCU = coronary care unit, STICU = surgical/trauma ICU, MICU = Medical ICU, CVICU = Cardiovascular ICU, OR = operating room, ED = emergency department, Ang-2 = angiotensin II, MAP = mean arterial pressure, IQR = interquartile range, APACHE = Acute Physiology and Chronic Health Evaluation.

Angiotensin II was initiated across a broad variety of locations such as the medical, cardiovascular, and surgical intensive care units. The median number of vasopressors prior to Ang-2 administration was 3 (IQR 2–3). The median Acute Physiology and Chronic Health Evaluation (APACHE) IV score was 90 (IQR 71–121), and 68.5% of patients were on ≥ 3 vasopressors prior to Ang-2 administration. Hydrocortisone (69.8%) was the most common rescue agent administered, followed by ascorbic acid (49.6%), hydroxocobalamin (7.9%), and methylene blue (5.8%). Mechanical ventilation was required in 94.1% of patients.

The initial median Ang-2 dose was 10 ng/kg/min (IQR 10–20) and the median maximum dose was 40 ng/kg/min (IQR 29–80). At the time of Ang-2 initiation, the median MAP was 62 mmHg (IQR 57–70), while the median NEpi-equivalent dose was 0.55 μg/kg/min (IQR 0.38–0.82) (Table 2). Three hours after Ang-2 administration, the median MAP increased to 73 mmHg (65–79) and the median NEpi-equivalent dose decreased to 0.39 μg/kg/min (IQR 0.16–0.64) (Figure 1). The mean difference in MAP at 3 hours was 9.3 mmHg (95% CI 6.4–12.1 mmHg, p < 0.001) and the mean difference in NEpi-equivalent dosing at 3 hours was 0.16 μg/kg/min (95% CI 0.10–0.22, p < 0.001). The median time to hemodynamic response, which was defined as an increase in MAP ≥ 65 or a MAP increase ≥ 10 mmHg, was 16 minutes (IQR 5–60 min) and the median duration of Ang-2 therapy was 25.0 hours (9.9–45.9). In patients with hemodynamic response to Ang-2 by hour 3, this target was reached within 15, 30, and 60 minutes in 49.2%, 63.3%, and 82.0% of patients, respectively (Table 3).

Table 2.

Change in Hemodynamics After Angiotensin II Administration.

Variable	Hour 0	Hour 3	Mean difference (95% CI)	P
Hemodynamic Data
MAP, mmHg	62 (57–70)	73 (65–79)	9.3 (6.4–12.1)	<0.001
NEpi-equivalent dose, μg/kg/min	0.55 (0.38–0.82)	0.39 (0.16–0.64)	0.16 (0.10–0.22)	<0.001

Continuous variables are represented as median (interquartile range). Hour 0 refers to measurement of variable immediately prior to Ang-2 administration. Hour 3 refers to measurement of variable 3 hours after Ang-2 administration. Wilcoxon signed-rank test performed for continuous variables.

P-values are adjusted for multiple comparisons utilizing Bonferroni correction.

Abbreviations: kg = kilograms, MAP = mean arterial pressure, CI = confidence interval, NEpi = norepinephrine.

Figure 1.

Hemodynamic parameters at hour 0 and hour 3 of therapy. Shown are MAP (upper panel) and NEpi-equivalent dose (lower panel) at hours 0 and 3 of Ang-2 treatment. The mean differences in hemodynamic parameters were calculated with a 95% confidence interval (CI) and compared with a Wilcoxon signed-rank test. There was a significant improvement in MAP and NEpi-equivalent dose after the addition of angiotensin II, p < 0.001 for both outcomes. In the box plots, the horizontal line inside each box indicates the median, the top and bottom of the box indicate the interquartile range, the I bars indicate the 5th and 95th percentiles, and the circles indicate outliers. The dashed line indicates a MAP of 65 mmHg. Abbreviations: MAP = mean arterial pressure, NEpi = norepinephrine, mmHg = millimeters of mercury, mcg = micrograms, NEpi = norepinephrine, ng = nanograms, mL = milliliters, hr = hour.

Table 3.

Angiotensin II Characteristics, Hemodynamic Effects, and Adverse Events.

Angiotensin II characteristics
Initial Ang-2 dose, ng/kg/min	10 (10–20)
Maximum Ang-2 dose, ng/kg/min	40 (29–80)
Duration of Ang-2 therapy, hours	25.0 (9.9–45.9)
Hemodynamic effects
Time to MAP ≥65, min	16 (5–60)
% Reduction in NEpi-equivalents by Hour 3	17.4 (0.0–57.7)
MAP ≥ 65 or increase by 10 mmHg by Hour 3, %	80.1%
When target MAP achieved, % response within
15 minutes	49.2%
30 minutes	63.3%
60 minutes	82.0%
Renal outcomes
Serum Creatinine at baseline, mg/dL	2.0 (1.3–3.0)
AKI prior to Ang-2, %	75.0%
RRT during hospitalization, %	47.9%
RRT initiated before Ang-2	97.1%
RRT initiated after Ang-2	2.9%
Liberation from RRT during hospitalization, %	64.3%
Total days on RRT, days	5 (2–8)
Thromboembolic data
VTEs, %	3.1%
Grade 1	1.9%
Grade 2	1.2%
Grade 3–5	0.0%
DVT prophylaxis utilization, %	75.4%

Abbreviations: Ang-2 = angiotensin II, ng = nanograms, kg = kilograms, min = minute, MAP = mean arterial pressure, NEpi = norepinephrine, VTE = venous thromboembolic event, DVT = deep vein thrombosis, RRT = renal replacement therapy.

Responders to Ang-2 had a higher MAP (10.0 mmHg responders vs. −4.5 non-responders, p < 0.001) and lower NED (−0.12 μg/kg/min responders vs. 0.00 non-responders, p = 0.02) at hour 3 than those who were non-responders. When stratifying patients by the NED dose prior to the initiation of Ang-2, those with a NED dose < 0.2 μg/kg/min had a significantly greater reduction in NED by hour 3 (71.5% reduction in those below 0.2 μg/kg/min threshold vs. 26.3% increase in those above threshold, mean difference −97.7%, 95% CI −171.7% to −23.8%, p = 0.01). Similarly, when utilizing a threshold < 0.3 μg/kg/min, this difference was also significant, albeit smaller in magnitude than the 0.2 μg/kg/min threshold (42.4% NED reduction in those below 0.3 μg/kg/min threshold vs. +25.9% increase in those above threshold, mean difference −68.3%, 95% CI −133.5 to −3.0%, p = 0.04). However, when utilizing a threshold < 0.4 μg/kg/min or < 0.5 μg/kg/min, the relative reduction in NED by hour 3 was no longer significant (p = 0.08 and p = 0.22, respectively). Finally, when stratifying patients by the number of vasopressors prior to the initiation of Ang-2, those patients on ≤ 3 vasopressors had a significantly greater improvement in relative NED reduction than those patients on >3 vasopressors (−28.2% ≤ 3 vasopressors vs. +28.2% > 3 vasopressors, mean difference 56.4%, 95% CI 1.52–111.3, p = 0.04).

The median serum creatinine prior to Ang-2 administration was 2.0 mg/dL (IQR 1.3–3.0). Acute kidney injury was present in 75.0% of patients, and RRT was utilized in 47.9% of patients prior to Ang-2 administration. In those that required RRT, it was initiated 97.1% of the time before Ang-2 administration and 2.9% of the time after Ang-2 administration. During the hospitalization, patients required RRT for a median of 5 days (IQR 2–8), and 64.3% patients on RRT were liberated from it prior to discharge from the hospital.

The median time to first and second vasopressor discontinuation after Ang-2 administration was 8.0 hours (IQR 0.5–21.5) and 19.8 hours (3.1–49.8), respectively (Table 4). When assessing the order of vasopressor discontinuation, the first vasopressor discontinued was most commonly phenylephrine (25.8%), followed by vasopressin (24.2%) and NEpi (23.4%) (Table 4). The second discontinued vasopressor was most commonly NEpi (34.5%) and vasopressin (30.2%), followed by epinephrine (12.9%) and phenylephrine (12.1%).

Table 4.

Characteristics of Vasopressor Discontinuation.

Variable	1st	2nd	3rd	4th
Order of vasopressor removal
Norepinephrine	23.4%	34.5%	34.9%	48.3%
Epinephrine	21.0%	12.9%	16.9%	6.9%
Phenylephrine	25.8%	12.1%	9.6%	3.4%
Vasopressin	24.2%	30.2%	28.9%	37.9%
Dopamine	0.0%	0.9%	0.0%	3.4%
Dobutamine	2.4%	0.9%	0.0%	0.0%
Ang-2	3.2%	8.6%	9.6%	0.0%
Time to vasopressor removal
Hours	8.0 (0.5–21.5)	19.8 (3.1–49.8)	49 (20.7–103)	80 (24.2–152.5)
Hemodynamics after vasopressor removal
MAP, mmHg	75 (68–85)	75 (70–81)	78 (71–83)	72 (69–78)

Continuous variables are represented as median (interquartile range).

Abbreviations: MAP = mean arterial pressure.

DVT prophylaxis was utilized in 75.4% of patients. Venous thromboembolism within 28 days of Ang-2 administration occurred in 3.1% of patients. Three DVTs occurred in the upper extremity and 2 occurred in the lower extremity. All DVTs were classified as Grade 1–2 (mild-moderate) severity.

Discussion

Our study suggests that Ang-2 is an effective vasopressor and catecholamine-sparing agent. We identified significant increases in 3-hour MAP and significant improvements in background NEpi-equivalent dose after Ang-2 was initiated. These results are statistically and clinically important because, contemporary evidence examining the use of human Ang-2 is mainly from the ATHOS-3 trial and its subsequent post-hoc analyses.^3,9–16 Our study includes patients with a wide variety of shock states, including cardiogenic and hypovolemic shock, which were not studied in ATHOS-3.

Patients in our study were critically ill, with 68.5% of patients receiving ≥3 vasopressors prior to Ang-2 administration. A severe level of illness is also reflected by a median APACHE IV score of 90 for patients in our cohort. The duration of Ang-2 therapy was relatively short, 25.0 hours, in our study. Our data are consistent with ATHOS-3, in that Ang-2 increased MAP ≥65 or by 10 mmHg in a significant proportion of patients. When target MAP was reached, it occurred rapidly, with 49.2% of patients reaching it within 15 minutes, 63.3% within 30 minutes, and 82.0% within 60 minutes. These effects are consistent with the drug’s short half-life from plasma metabolism by aminopeptidase A and angiotensin-converting enzyme 2.¹

Our study additionally demonstrates that Ang-2 is an effective catecholamine-sparing agent, with a mean reduction in NEpi-equivalent dose by 0.16 μg/kg/min at hour 3 of therapy. The reduction in NEpi-equivalent dose was clinically and statistically significant. This finding is important because among patients in shock requiring high dose catecholamines, the mortality has been reported to be as high as 83%.²⁰ Sviri and colleagues found that patients receiving ≥40 μg/min of NEpi or epinephrine had an ICU mortality of 84%, compared to a mortality of 50% in the group receiving <40 μg/min.²¹ After multivariate logistic regression analysis, high-dose catecholamines were found to be an independent predictor of ICU mortality, and patients in this cohort were 5.1 times more likely to have ICU mortality than patients on low-dose catecholamines (OR 5.1, 95% CI 2.02–12.9, p = 0.001).²¹ Adverse events such as digital ischemia, limb ischemia, and mesenteric ischemia are not uncommon when treating shock with high-dose catecholamines.²² Brown and colleagues found that 7.9% of all patients with shock on high-dose vasopressors developed digital or limb necrosis, while Schmittinger and colleagues demonstrated that the number of catecholamines and duration of therapy was significantly associated with adverse cardiac events.^20,23 The catecholamine-sparing effects of non-catecholamine vasopressors may have benefit in reducing mortality and adverse effects associated with high-dose catecholamines.^24,25 This relationship, however, is highly impacted by severity of illness, and the direct impact of catecholamine-sparing agents on mortality from shock is unclear. Future prospective research should further evaluate the relationship between Ang-2, catecholamine requirements, and clinical outcomes.

Angiotensin II is not yet part of the Surviving Sepsis Guidelines for the treatment of septic shock, and there are many questions as to when it should be initiated in the treatment algorithm.¹⁸ In ATHOS-3, patients received Ang-2 after they required a NED dose ≥ 0.2 μg/kg/min for greater than 6 hours, and it was mostly utilized as an early third line agent.³ Since then, there has been wide variability in its use, and case reports have described the initiation of Ang-2 anywhere between 0.16 to 3.0 μg/kg/min, and as early as a second-line agent to as late as a fifth-line agent.^{14,16,13,26,27} In our study, patients with a NED dose <0.2 μg/kg/min or <0.3 μg/kg/min at the time of Ang-2 administration benefited from a greater relative reduction in NED by hour 3, and this reduction was greater in magnitude at the 0.2 μg/kg/min threshold. This benefit was not statistically significant at a NED threshold of <0.4 or <0.5 μg/kg/min. Finally, those on ≤ 3 vasopressors at the time of Ang-2 administration had a statistically significant reduction in NED, whereas this was not seen in those on >3 vasopressors. Taken together, our findings suggest that consistent with the ATHOS-3 data, that Ang-2 should be utilized early in the course of shock, before the initiation of the fourth-line vasopressor, and before the NED dose exceeds 0.3 μg/kg/min, but ideally before it exceeds 0.2 μg/kg/min.

The increased incidence of VTEs in ATHOS-3 has been highly scrutinized. One systematic review of 31,000 patients in 1,126 studies examined the safety of Ang-2 and found that adverse events attributable to Ang-2 were rare and most frequently consisted of headache, chest pressure, nausea, and dizziness.²⁸ In ATHOS-3, the incidence of VTEs was 12.9% in the Ang-2 group and 5.1% in the standard of care (SOC) group.³ This analysis examined both clinically significant and clinically non-significant VTEs, and not all study centers utilized DVT prophylaxis as standard protocol in the trial. When examining only the clinically significant VTEs, the difference between the 2 groups was not statistically significant (1.8% Ang-2 vs. 0.0% SOC, p > 0.05).³ In one meta-analysis of 1783 critically ill patients who did not receive Ang-2, the overall DVT rate was 12.7%.²⁹ Although it was not possible to perform direct statistical comparisons with these studies, the VTE incidence in our study was 3.1%, which is consistent with the 5.1% rate seen in the control group of the ATHOS-3 trial, and less than the 12.7% rate observed in a meta-analysis of 1,783 ICU patients.³,²⁹ None of the VTEs in this study were clinically severe based on the National Cancer Institute criteria.¹⁹

Our study has several important limitations. This is a retrospective, observational study, and therefore, it is prone to confounding and bias. Data collection was dependent on accurate documentation in the medical record and on consistent extraction of that data from multiple investigators across several sites. Our dataset did not contain data on fluid balance or use of mechanical support in cardiogenic shock, which may skew our results. The sample size is also relatively small and is limited by the available data. Because of this, adverse event data may be particularly prone to under-reporting. Although this was a multicenter study, the heterogeneity and volume of patients limits our ability to perform subgroup analyses and to definitively draw conclusions about mortality in certain populations where there could be incremental benefit.^9–12 Furthermore, our study did not have a separate control group for comparison. Instead each patient served as their own internal control, and we observed hemodynamics and vasopressor dose before and after treatment with Ang-2. It is possible that without Ang-2 treatment, the NED would have decreased over time due to other factors. Nevertheless, our study remains one of the largest studies of Ang-2 since publication of the ATHOS-3 trial and provides continued evidence that Ang-2 is an effective treatment for vasodilatory shock.

Conclusion

In this multi-center observational cohort study, we demonstrated that Ang-2 significantly increased MAP by hour 3, with a median time to effect of 16 minutes. Furthermore, Ang-2 resulted in a significant reduction in NED. Our data indicates that Ang-2 should be utilized early in the course of shock, before the initiation of the fourth-line vasopressor, before the NED dose exceeds 0.3 μg/kg/min, and ideally before it exceeds 0.2 μg/kg/min. The incidence of VTEs was 3.1%, which is comparable or lower than rates previously reported in critically ill patients.²⁹ Given the critical importance of adequate blood pressure to organ perfusion, Ang-2 should be considered when target MAP cannot be achieved with conventional vasopressors.