Continuous Hemodynamic Response to Angiotensin II in Critically Ill Pediatric Patients: A Single Center Cohort Study

Introduction

Pediatric sepsis occurs worldwide in 1.2 million children annually. Mortality can be as high as 50% and is most often attributed to refractory shock and/or multiorgan dysfunction. Current pediatric guidelines recommend catecholamine agents as first-line treatment for fluid refractory shock.¹ Owing to receptor downregulation and endogenous catecholamine metabolism, patients often require increasing amounts of catecholamines to maintain hemodynamic stability and adequate end-organ perfusion.² High doses and extended durations of catecholamine agents can lead to serious adverse effects including peripheral ischemia, myocardial cellular apoptosis, and arrhythmia development.³ The addition of vasopressin is suggested if patients require high doses or continued titration of catecholamines.¹ Additional non-catecholamine agents used in clinical practice, including stress-dose hydrocortisone and methylene blue, lack reliable efficacy data to be strongly recommended in pediatric guidelines and remain last-line agents.^4–6 Disadvantages of high-dose catecholamine and vasopressin use, combined with a paucity of literature regarding third-line agents, necessitate additional research for an efficacious vasopressor with a mechanism of action that augments guideline recommendations.

Angiotensin II (AT-II) is a novel non-catecholamine vasopressor that binds to peripheral AT-II receptors resulting in vasoconstriction. In adults, AT-II increased mean arterial blood pressure (MABP) and reduced norepinephrine vasopressor equivalents at hour 3 compared with placebo leading to US Food and Drug Administration (FDA) approval in adults for use in septic or other distributive shock in 2017.⁷ Angiotensin II similarly reduced vasopressor requirements while maintaining goal MABP in a small number of pediatric patients; however, data are limited to descriptive case reports and a single center case series that rely on intermittent blood pressure evaluation.^8–10 At St. Louis Children’s Hospital, AT-II is used in pediatric patients with vasodilatory shock who require multimodal vasopressor therapy and who cannot meet or maintain their MABP goal. Our institution also uses data management software that captures continuously monitored vital sign data from patients admitted to an intensive care unit (ICU).

In this study, we aimed to use automated, high-fidelity hemodynamic data and vasopressor requirements to study the efficacy of AT-II in a pediatric population with vasodilatory shock requiring multimodal vasopressor therapy. Also, we aimed to identify patient and clinical characteristics of patients who responded or did not respond to AT-II therapy. Finally, we aimed to report AT-II dosing strategy and adverse events following AT-II initiation.

Materials and Methods

Patients admitted to the St. Louis Children’s Hospital Pediatric Intensive Care Unit (PICU) who received AT-II between June 2017 and November 2022 were identified by using the hospital electronic medical record (EMR). The availability of corresponding hemodynamic data in the Etiometry T3 Quality Improvement System (Etiometry Platform, Etiometry Inc, Boston MA) during AT-II administration was evaluated. Etiometry T3 is an FDA-cleared data management software integrated into the EMR system that captures and archives prospective ICU clinical data from multiple sources in near real-time. Patients were included if they received AT-II for at least 3 hours and had hemodynamic data from arterial access archived in Etiometry T3 for the duration of the AT-II infusion. Patients were excluded if they received AT-II for less than 3 hours or did not have hemodynamic data available for the duration of the infusion. This study was approved by the Institutional Review Board at Washington University in St. Louis.

Baseline patient characteristics and clinical data were collected through retrospective chart review. Baseline demographics included patient age, sex, weight, severity of illness, and cause of vasodilatory shock. Race and ethnicity were not collected owing to the impact of potential socioeconomic, health care, and social constructs on the incidence and outcomes of pediatric septic shock and the inability to account for these confounders.¹¹ Severity of illness at PICU admission was quantified by calculating the Pediatric Risk of Mortality (PRISM-III) score.¹² The Pediatric Logistic Organ Dysfunction–2 (PELOD-2) score was calculated at PICU admission, prior to AT-II initiation, and at AT-II discontinuation.¹³ Additional data collected prior to AT-II initiation included stress-dose hydrocortisone use, need for mechanical ventilation, need for renal replacement therapy (RRT), left ventricular fractional shortening percentage, and serum lactate concentration. The use of venous thromboembolism (VTE) prophylaxis or documented contraindications, the need for extracorporeal membrane oxygenation (ECMO) during the PICU admission, acute kidney injury (AKI), acute respiratory distress syndrome (ARDS), and mortality during AT-II infusion and at 30 days were captured. Acute kidney injury was assessed 48 hours prior to AT-II initiation through 24 hours post AT-II discontinuation and categorized by worst AKI stage using standard pediatric Kidney Disease: Improving Global Outcomes (KDIGO) definitions.¹⁴ ARDS severity was calculated from data within 6 hours prior to AT-II initiation as outlined by the Pediatric Acute Lung Injury Consensus Conference Group.¹⁵ Patients were assessed for adverse events including cardiac arrest, cardiac dysfunction, arrhythmia, thrombosis, and peripheral ischemia through review of daily progress notes during ICU admission. Imaging studies were reviewed for any patient with a documented thromboembolic complication to determine presence before or after AT-II administration.

Angiotensin II first became available at our institution for treatment of refractory shock in 2017, with use at the discretion of the attending intensivist. In 2020, an institutional protocol was implemented to guide AT-II use. Criteria included clinical features of vasodilatory shock in the setting of adequate volume resuscitation and cardiac output and the inability to obtain or maintain desired MABP despite high-dose multimodal vasopressor therapy (minimum criteria of ≥0.2 mcg/kg/min of epinephrine or norepinephrine and ≥0.5 milliunits/kg/min or 40 milliunits/min of vasopressin). The recommended starting dose for AT-II was 5 ng/kg/min with titrations every 5 minutes to a maximum dose of 80 ng/kg/min based on a blood pressure goal determined by the attending intensivist.

Information on AT-II and other vasopressor dosing and duration was collected via documentation in the electronic medication administration record and the nursing critical care flowsheets to create a single vasopressor titration timeline. Weaning of AT-II and other vasopressors was at the discretion of the attending intensivist. Non–AT-II vasopressor support was quantified into a single vasoactive-inotropic score (VIS).^16,17 The VIS was calculated as a continuous variable and data were imputed from the most recent documented dose for each vasopressor. Baseline VIS reflects the median VIS in the 10 minutes immediately prior to AT-II initiation. High-fidelity hemodynamic data were accessed via the Etiometry T3 platform, which prospectively collected second-by-second blood pressure data via arterial line. All blood pressure data analyzed were obtained from an invasive hemodynamic monitoring device. Data for MABP were analyzed as the median values per minute. Baseline MABP was the median MABP during the 10 minutes immediately prior to the start of the AT-II infusion. To account for the variance in MABP goals based on patient age, the difference between an evaluated MABP and a patient’s baseline MABP was calculated (ΔMABP).

The primary objective of our study was to assess acute hemodynamic response after initiation of AT-II. A response was defined as either an increase in ΔMABP by ≥10 mm Hg and/or 15% at 3 hours after AT-II initiation with a stable VIS (change of ≤10) or a reduction in VIS of >10 at 3 hours after AT-II initiation.⁷ A change in VIS (ΔVIS) of >10 was used because it corresponds to a norepinephrine equivalent dose (NED) change of >0.1 mcg/kg/min. The 3-hour ΔMABP and VIS were calculated by using the median value over a 3-minute span at hour 3 of the AT-II infusion. Demographic and clinical characteristics were compared between AT-II responders and nonresponders. Secondary analyses included impact of AT-II infusions on PELOD-2 scores over time, description of AT-II dosing, use of VTE prophylaxis, and assessment for potential adverse effects.

Descriptive statistics were used to summarize demographic data, baseline clinical characteristics, vasopressor use, AT-II dosing, VTE prophylaxis, and adverse effects. Univariate comparisons of demographic and clinical characteristics between responders and nonresponders were performed by using the Mann-Whitney U test and Fisher’s exact test for non-normally distributed continuous and categorical data, respectively. The Wilcoxon Rank Sum test was used to compare paired MABP and VIS values between baseline and 3 hours and PELOD-2 between AT-II initiation and discontinuation. Continuous ΔVIS and MABP values are depicted as smoothed conditional means modeled using locally estimately scatterplot smoothing (LOESS) non-linear regression with the line of best fit and the 95% confidence interval estimated. Statistical tests were considered significant where p < 0.05. Statistical analysis was performed with SPSS version 25 (Armonk, NY) and R version 4.2.2 (Vienna, Austria) including the ggplot2 package.

Results

A total of 25 pediatric patients were identified as receiving an AT-II infusion during the study period. Six patients were excluded for lack of continuous hemodynamic data, 4 patients had incomplete or inaccessible continuous hemodynamic data, and 1 patient had AT-II discontinued prior to the 3-hour response evaluation. Fourteen patients had complete continuous hemodynamic data suitable for the primary analysis. The median age and weight were 11.6 years (range, 13 days–16.8 years) and 34.6 kg (range, 2.9–76.8), respectively. Ten patients (71%) received AT-II for the management of septic shock and received stress-dose hydrocortisone prior to AT-II initiation. One patient (7%) received 1 vasopressor, 6 (43%) received 2 vasopressors, and 7 (50%) received 3 vasopressors prior to AT-II initiation; the median baseline VIS of 50 reflected a correspondingly high vasopressor dose requirement. Additional demographic and baseline clinical characteristics are presented in Table 1.

Table 1.Baseline Clinical Characteristics^*

View Fullscreen

The median duration of AT-II infusion for the cohort was 32 hours (range, 3.1–72.4). The median starting and maximum dose of AT-II were 5 ng/kg/min (range, 2.5–10) and 80 ng/kg/min (range, 20–80), respectively. The median time to maximum dose was 58 minutes (range, 17–434). At 3 hours, the median ΔMABP was significantly higher than baseline by 7 mm Hg (range, −5 to 40) (p = 0.04) and the VIS was significantly decreased at −4.5 (range, −100 to 15) (p = 0.01). A total of 10 patients (71%) were categorized as responders. Figure 1 shows the trend in VIS and percent change in MABP between groups. Of the responding patients, 5 (50%) met response criteria based on a decrease in VIS of >10 and 6 (60%) met response criteria based on a positive ΔMABP of >10 mm Hg and/or 15% increase with a stable VIS (1 patient met both criteria). As expected, change in VIS was significantly higher in responders at 3 hours of therapy (−12 vs −1; p = 0.02). Supplemental Figure S1 depicts the best-fit lines and 95% CI modeled by using LOESS nonlinear regression for the complete treatment period. Responders were significantly older (p = 0.002) and had a significantly higher baseline VIS (p = 0.008) than nonresponders (Table 2). Despite similar baseline PRISM-III and PELOD-2 scores, nonresponders had over a three-fold higher mortality rate during the AT-II infusion and at 30 days; however, these differences were not statistically significant (Table 2). Due to this higher mortality rate, an assessment of PELOD-2 scores during AT-II administration could only be performed in the responder group. Of the 8 patients in the responder group who survived during the AT-II infusion (1 died after stopping the AT-II, but prior to 30 days), there was a significant decrease in the PELOD-2 score from AT-II initiation to AT-II discontinuation (median, 17; range, 4–23 vs median, 11; range, 2–17; p = 0.01). All other characteristics were similar between responders and nonresponders (Table 2).

Table 2.Patient Characteristics Based on Response to AT-II^*

View Fullscreen

Two patients (14%) had a documented thrombosis, but both were identified prior to AT-II initiation. One patient (7%) experienced ischemia of the right lower extremity, which did not result in AT-II discontinuation. None of the patients in our cohort received pharmacologic VTE prophylaxis during the AT-II infusion. Three patients (21%) were already receiving therapeutic anticoagulation with either bivalirudin or unfractionated heparin. Ten patients (71%) had contraindications to anticoagulation, including 8 patients (57%) with thrombocytopenia and 2 patients (14%) with coagulopathy. Non-pharmacologic VTE prophylaxis with sequential compression devices were used in 2 patients (14%). No other adverse effects were documented.

Discussion

To date, this is the only study to describe response to AT-II, using high-fidelity hemodynamic trends in a pediatric population. Despite the high degree of critical illness within this cohort, treatment with AT-II resulted in a positive acute hemodynamic response in most patients. Classification of response was equally distributed between patients who had a positive ΔMABP or a decrease in VIS at hour 3 of the AT-II infusion. Baseline VIS was higher in those patients with an acute hemodynamic response to AT-II initiation. Our study is also the first to assess severity of illness in pediatric patients across the duration of the AT-II infusion, showing a significant decrease in PELOD-2 score in responders, indicating improved organ function and reduced mortality risk. Documented adverse effects were infrequent and did not require discontinuation of AT-II.

Angiotensin II has robust evidence evaluating efficacy, safety, and predictors of response in the adult population. Angiotensin II effectively increased MABP compared with placebo in adults with vasodilatory shock who were unresponsive to high-dose multimodal vasopressor therapy in the ATHOS-3 trial.⁷ Post hoc analyses of ATHOS-3 documented a more profound response to AT-II therapy in patients with AKI requiring RRT and decreased mortality with initiation of AT-II at lower doses of other vasopressors.^7,18,19 A subsequent retrospective, multicenter, adult study demonstrated a significant association between favorable hemodynamic response and lower baseline lactate concentrations.²⁰ Informed by these trials, we explored clinical factors predictive of response to AT-II in pediatric patients. We were unable to reproduce these findings in our small pediatric cohort. However, we evaluated several pediatric-specific characteristics, including age, weight, and left ventricular fractional shortening percentage, and found that responders to AT-II were significantly older with correspondingly higher weights (p = 0.002 and p = 0.004, respectively). We are unable to attribute this difference in response to myocardial maturity because all patients in our cohort with available data (n = 10) had normal left ventricular fractional shortening percentages (Table 2). Using a less specific definition of response, Tezel et al¹⁰ similarly found no difference in clinical characteristics, including need for ECMO, RRT, and steroid use, between survivors and those who died after AT-II infusion. Interestingly, although our time to AT-II initiation was less in the responder group, responders had a significantly higher VIS (p = 0.008) than nonresponders. Tezel et al¹⁰ similarly showed a numerically lower time on vasopressors in pediatric patients who survived but with a correspondingly lower NED; however, in both evaluations this difference was not statistically significant.

Data supporting the efficacy and safety of AT-II in the pediatric population are limited to small case series and a single center cohort study.^8–10 Although these studies demonstrated an increase in MABP and/or decrease in requirement for vasopressor support with the use of AT-II, definition of hemodynamic response was subjectively determined and results are limited by intermittent blood pressure assessments collected retrospectively from the EMR. In our study, we used high-fidelity, continuous hemodynamic assessment with retrospective clinical information to categorize response to AT-II in a critically ill pediatric population. We modified previously used adult response definitions to account for criteria not directly applicable to pediatric patients, including varying blood pressure goals for age and use of a validated pediatric vasopressor equivalence score. More specifically, we adapted the definition of blood pressure response used in the ATHOS-3 trial (increase in MABP from baseline of at least 10 mm Hg or an increase to at least 75 mm Hg) to account for young pediatric patients with lower goal blood pressures (increase in MABP from baseline of at least 10 mm Hg or 15%). We defined a clinically meaningful change in VIS of >10 to correlate with a NED change of >0.1 mcg/kg/min, which was the change in background vasopressor criteria used in ATHOS-3 to continue use of study medication.⁷ Our definition of response, using both a ΔMABP/% change and ΔVIS at 3 hours, may provide better feedback than either blood pressure response or change in vasopressor requirement alone for clinicians and researchers evaluating AT-II in pediatrics. Consideration of both variables in the first 3 hours of AT-II administration, in addition to patient-specific characteristics, may help with intensivist assessment of response and decisions regarding continuation of treatment. Use of similar criteria should be considered for AT-II response assessment in future prospective studies. Additionally, use of automated, high-fidelity, continuous hemodynamic data obtained from invasive monitoring devices improved precision in identification of response to AT-II in the pediatric population (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

The largest pediatric cohort study (Tezel et al¹⁰) described the use of AT-II in 23 critically ill pediatric patients with catecholamine-resistant vasodilatory shock. Guidelines for use of AT-II were similar between our institutions, resulting in patient populations with comparable PELOD-2 scores at AT-II initiation, rates of invasive mechanical ventilation, use of steroids, and a primary etiology of sepsis. Tezel et al¹⁰ documented an increase in median MABP of 6.5 mm Hg and decrease in the median NED by 18%, 3 hours after initiation of AT-II. We saw a comparatively positive median ΔMABP in our entire cohort of 7 mm Hg at 3 hours with a smaller decrease in median VIS of 9%. While both the VIS in our cohort and the NED in the cohort presented by Tezel et al¹⁰ represent a high vasoactive burden on their respective scales, it is difficult to directly compare NED and VIS given differences in their calculations. Most notably the weight of vasopressin varies drastically; whereas VIS escalates consistently with weight-based titrations common to pediatric practice, NED assumes flat dosing based on adult data and practice and application of this ratio to titratable doses increases the influence of vasopressin on the score.²¹ For this reason and given current data supporting validated use in the pediatric population, we opted to use VIS in the current report.^16,17 Similar to previous pediatric cohorts, mortality in the current study was high; however, this is unsurprising given our use criteria, baseline vasopressor requirements, and PELOD-2 score at AT-II initiation. While the difference in patient ages between responders and nonresponders in this cohort is concerning, it is important to recognize the known worse outcomes in younger patients with septic shock, which could account for the mortality trends observed between groups.²² Furthermore, this difference may be attributed to ontogenetic changes in the renin-angiotensin-aldosterone system. Endogenous AT-II levels are generally elevated in early infancy, but they decrease with age. This age-related shift may present a potential for therapeutic use of exogenous AT-II, though current data on endogenous AT-II concentrations in infants and children show considerable variability.²³ We believe that age-related differences may have pathophysiologic feasibility and demonstrate the necessity of future studies of the very young. Safety data for AT-II also exist primarily in adult literature indicating a potential for tachyarrhythmias, peripheral ischemia, and deep-vein thrombosis.⁷ Among 23 pediatric patients in a recent case series, a single occurrence of digital ischemia and a peripherally inserted central catheter associated thrombus were reported.¹⁰ These reports, in addition to our single report of peripheral ischemia, are unable to establish direct causality given their retrospective design, high baseline vasoactive burden, and inadequate sample size to reliably detect rare adverse effects.

Additional limitations of this cohort study should be noted. Use of medication guidance criteria resulted in many patients receiving AT-II as a last-line therapy as demonstrated by an increasing PELOD-2 score and high VIS at AT-II initiation. This prevents evaluation of a defined place in treatment for AT-II, may confound the number of responders and mortality rate reported, and limits applicability of our results to less critically ill patients. Furthermore, the individualized titration practices of AT-II by each physician may have influenced the response rate, although 79% of patients reached their maximum AT-II dose prior to the 3-hour efficacy mark and there was no difference in time to achieve maximum AT-II dose between responders and nonresponders (Table 2). Additionally, the retrospective design of this study limited our ability to evaluate adequacy of the volume resuscitation component of the use criteria because it was determined at the discretion of the attending physician. Although we collated vasopressor dose titrations from multiple documentation sources, vasopressor dosing was not collected from an infusion pump integrated with the electronic health record, and therefore calculated VIS scores are limited by manual charting. Similarly, adverse effect reporting relied on physician documentation and association of adverse effects with AT-II was highly subjective.

Conclusions

Angiotensin II has a potential role in the management of pediatric patients with vasodilatory shock based on improvement in MABP observed in adult studies and emerging pediatric data. This study used high-fidelity, continuous hemodynamic monitoring to contribute additional data on the acute hemodynamic response to AT-II in pediatric patients with fluid and vasopressor refractory shock, documenting a positive response in 71% of patients corresponding with a significant decrease in the PELOD-2 scores. While a difference in age, weight, and baseline VIS were noted in our cohort, patient-specific predictors of response to AT-II warrant further exploration.