FDA Approval of Angiotensin II for the Treatment of Hypotension in Adults With Distributive Shock

Fortunato Senatore; Gowraganahalli Jagadeesh; Martin Rose; Venkateswaran C. Pillai; Sudharshan Hariharan; Ququan Liu; McDowell Tzu-Yun; Mohan K. Sapru; Mary Ross Southworth; Norman Stockbridge


Am J Cardiovasc Drugs. 2019;19(1):11-20. 

In This Article

Study Outcomes

Baseline Characteristics of Subjects in the ATHOS-3 Trial

A total of 404 patients were screened. Of these 60 were excluded (54 did not meet eligibility criteria, three withdrew consent, one died, one moved to hospice, and one for reasons unknown). The remaining 344 patients were randomized 1:1 (i.e., 172 per arm) and were included in the intent-to-treat (ITT) population. Of these, 13 in the placebo arm and 10 in the angiotensin II arm did not receive study drug. Also, one subject initially randomized to the placebo arm was re-randomized to angiotensin II, and was considered an angiotensin II arm subject for analysis. Thus, there were 158 and 163 subjects in the placebo and angiotensin II arms, respectively, in the mITT population. The baseline characteristics between subjects in the placebo and angiotensin II arms were similar (Table 2). The mITT population was 61% male, 81% white, and 10% black, had a mean age of 62 years (range 22–89 years), and 74% were from the USA/Canada. The mean baseline values were MAP 66 mmHg, APACHE II score 28, Model of End-stage Liver Disease (MELD) score 21, cardiovascular Sequential Organ Failure Assessment (SOFA) score 4.00, and total SOFA score 12.00. The values were evenly distributed between the arms of the trial. The APACHE II score is based on common physiological variables measured during the initial 24 h in the intensive care unit (ICU),[16] and it is correlated with mortality.[17] The SOFA score is sepsis related and evaluates major organ systems serially over time,[18] and it is also correlated with mortality.[19] The MELD score provides an estimate of mortality at 3 months from liver disease.[20]

The mean NED was 0.5 μg/kg/min during 6 h prior to randomization and continued during randomization. Vasopressin, whose doses were calculated as part of the NED, was administered to 70% of the ITT population in both arms during 6 h prior to randomization. The level of NED and the inclusion of vasopressin indicated a population with a significant mortality risk.

Dosing, Pharmacodynamics and Pharmacokinetics

During period 1 and period 2, the study drug titration scheme (Table 3) allowed for titration every 5 to every 15 min based upon the MAP relative to the target range.

During period 1, the NED remained constant. Following initiation of angiotensin II infusion at 20 ng/kg/min, at least 50% of patients achieved target MAP by 5 min, and the response was a function of the baseline MAP. As the blood pressure effects of angiotensin II reached steady state within 5 min of constant infusion, and doses were permitted to be up-titrated or down-titrated quickly to achieve target MAP, the starting dose of 20 ng/kg/min was considered reasonable for the intended patient population irrespective of the baseline MAP. The dosing pattern of angiotensin II showed an initial down-titration followed by a steady increase in the average angiotensin II infusion rate (Figure 1). The mean angiotensin II doses (ng/kg/min) at hours 0–1, 1–2, and 2–3 were 16, 24 and 32, respectively (Table 4). The mean placebo doses (ng/kg/min) in the same three time periods were 51, 91, and 113, respectively (Table 4). The maximum allowed dose during the first 3 h of treatment was 200 ng/kg/min, but none of the 21% of patients who received doses > 80 ng/kg/min achieved the target MAP. Therefore, labeling recommended the maximum dose of angiotensin II of 80 ng/kg/min during the first 3 h of infusion.

Figure 1.

Pattern of study drug dosing and MAP for hours 0–3. Data represents mean ± SEM. Ang II angiotensin II, MAP mean arterial pressure, SEM standard error of the mean

During period 2, the overall mean angiotensin II dose remained constant at 19 ng/kg/min. The NED decreased from a mean of 0.49 μg/kg/min to a mean of 0.43 μg/kg/min in the placebo group, and decreased from a mean of 0.43 μg/kg/min to a mean of 0.34 μg/kg/min in the angiotensin II group (Table 5). The greater decrease of the NED in the angiotensin II arm compared to the placebo arm in period 2 was likely a consequence of the angiotensin II effect, whereas the decline in both arms appeared to be associated with management and resolution of the underlying cause of distributive shock.

Serum levels of angiotensin II were similar between baseline and after 3 h of infusion. However, the serum levels of angiotensin I were reduced by approximately 40%. This could be possibly due to a negative bio-feedback mechanism, which could have suppressed renin secretion and reduced the formation of angiotensin I from renin. Therefore, angiotensin II levels remain unchanged, but angiotensin I levels were reduced in the circulation. The published literature reports were reviewed to provide additional pharmacokinetic information for angiotensin II. Briefly, angiotensin II reaches steady-state plasma concentrations in 5 min after IV infusion.[21] It is metabolized to angiotensin (1–7) and angiotensin 2–8 (angiotensin III) by aminopeptidase A and angiotensin converting enzyme 2 (ACE2),[7,22,23] respectively. The plasma half-life of angiotensin II is less than a minute.[24]

Review of Efficacy

In the angiotensin II arm, the MAP rose from a mean of 66 mmHg to a mean of 77 mmHg, corresponding to a 70% responder rate, on a mean dose of 21 (standard deviation [SD] 21) ng/kg/min. In the placebo arm, the MAP rose from a mean of 65 mmHg to a mean of 70 mmHg, corresponding to a 23% responder rate (odds ratio 8; 95% confidence interval [CI] 5–13, P < 0.00001) (Table 6). The results were consistent across all evaluated subgroups, including those based on age, gender, race, body mass index (BMI), or geographic region.

The secondary endpoint was the change from baseline in total and cardiovascular SOFA scores. The mean change from baseline at hour 3 in the cardiovascular SOFA score was − 0.01 (SD 0.11) in the placebo arm and − 0.13 (SD 0.55) in the angiotensin II arm. Given that the norepinephrine equivalent infusion rates were constant in this period, the cardiovascular SOFA score should have remained constant. The mean change from baseline at hour 48 in the cardiovascular SOFA score was − 1.28 (SD 1.65) in the placebo group and − 1.75 (SD 1.77) in the angiotensin II arm. These changes were attributable to standard-of-care response in both arms, as well as catecholamine sparing in the angiotensin II arm via replacement of catecholamines with angiotensin II that is not listed in the cardiovascular SOFA calculation. The mean change from baseline at hour 3 in the total SOFA score was + 0.47 (SD 1.82) in the placebo arm and + 0.81 (SD 1.93) in the angiotensin II arm. Similarly, the mean change from baseline at hour 48 in the total SOFA score was + 1.04 (SD 5.34) in the placebo arm and + 1.05 (SD 5.50) in the angiotensin II arm. The total SOFA score changes from baseline suggested no clinical benefit of angiotensin II over placebo in the ICU setting for distributive shock. Despite this, mortality over the course of the study was numerically lower in the angiotensin II arm (Figure 2).

Figure 2.

Kaplan–Meier plot of mortality to day 28. The cumulative mortality rate on day 28 was 54% in the placebo arm and 46% in Giapreza® arm. The hazard ratio (95% CI) was 0.78 (0.57–1.07). CI confidence interval

Review of Safety

The major safety concern was an excess of arterial and venous thromboembolic events: 21 subjects (12.9%) on angiotensin II versus eight subjects (5.1%) on placebo. The most common of these events was deep vein thrombosis (DVT), which occurred in seven versus no subjects. The standard care of patients with sepsis includes anticoagulant treatment for prophylaxis, although such prophylaxis was not required by the protocol and was not provided to all subjects in the ATHOS-3 trial. Our exploratory analyses showed that anticoagulation with heparin or related drugs was similar in both arms (~ 50%) prior to treatment and was more frequent in the angiotensin arm during the study (52% vs. 32%). Thus, the higher incidence of thromboembolic events in the angiotensin II arm could not be attributed to a lower use of prophylactic anticoagulation in that arm. Considering that the preclinical data indicate that angiotensin II has prothrombotic effects through various mechanisms,[25] the totality of evidence suggests that the risk of thromboembolic events could be drug related. Hence, a warning regarding thrombotic risk along with a recommendation for prophylactic anticoagulation was added in labeling of angiotensin II.

There were other minor safety signals observed in the ATHOS-3 trial; however, it was difficult to distinguish which adverse events were likely to be causally related to angiotensin II, because of the small safety database and the extremely ill population. Despite these limitations, the result of a favorable trend in mortality was reassuring of safety. Table 7 summarizes the adverse reactions at incidence of at least 4 and 1.5% higher in the angiotensin II arm than in the placebo arm.