Intravenous Aviptadil and Remdesivir for Treatment of COVID-19-associated Hypoxemic Respiratory Failure: Randomized, Placebo-controlled Trials

Abstract

Background:

Aviptadil (vasoactive intestinal peptide) is a lung-protective neuropeptide and remdesivir is a nucleotide prodrug of an adenosine analog, both of which may improve outcomes for COVID-19 patients with acute hypoxemic respiratory failure.

Methods:

Daily 12-hour infusions of aviptadil or placebo for three successive days and daily infusions of remdesivir or placebo for up to 10 successive days were studied using a master protocol for adults hospitalized for COVID-19 with acute hypoxemic respiratory failure. Participants could be randomized to both study treatments in a 2×2 factorial design or to just one of the agents. For both treatment comparisons, the primary outcome, assessed at Day 90, was a 6-category ordinal outcome ranging from return to home with liberation from supplemental oxygen (recovery) for ≥77 days to death. Mortality through Day 90 was a key secondary outcome. The independent data and safety monitoring board recommended stopping the aviptadil trial on May 25, 2022 for futility. On June 9, 2022, the sponsor stopped the trial of remdesivir due to slow enrollment.

Findings:

Four hundred seventy-one of the 640 planned participants were randomized to aviptadil/placebo, among whom 461 participants (231 allocated to aviptadil and 230 to placebo) received any infusion of blinded aviptadil. For the aviptadil/placebo comparison, at entry, 271 (59%) were on high flow nasal oxygen or non-invasive ventilation and 190 (41%) on mechanical ventilation or ECMO. The odds ratio (aviptadil/placebo) of being in a better outcome category at Day 90 was 1.11 (95% CI: 0.80–1.55; p=0.54). Cumulative mortality through Day 90 was 38% (n=86) for aviptadil and 36% (n=83) for placebo (HR 1.04; 95% CI: 0.77–1.41). A composite safety outcome including death, organ failure, serious infection, serious adverse events, and grade 3 or 4 adverse events through Day 5 occurred in 63% (n=146) of aviptadil and 56% (n=129) of placebo participants (OR 1.40; 95% CI: 0.94–2.08). Eighty-seven participants were randomized to remdesivir/placebo.

Interpretation:

Among patients with COVID-19-associated acute hypoxemic respiratory failure, aviptadil, compared to placebo, did not significantly improve clinical outcomes through Day 90. The smaller than planned sample size for the remdesivir trial does not permit definitive conclusions regarding safety or efficacy.

INTRODUCTION

SARS-CoV-2 infection may lead to acute hypoxemic respiratory failure, including acute respiratory distress syndrome (ARDS), which is often lethal. Even with adherence to current treatment guidelines, 60-day mortality remains 30–50% among COVID-19 patients with acute hypoxemic respiratory failure.^{1,2 3}

Aviptadil, the synthetic form of the 28-amino acid neuropeptide vasoactive intestinal peptide (VIP), has been proposed as a treatment for acute hypoxemic respiratory failure based on pleoiotropic lung-protective effects including increases in surfactant production, decreased cytopathy, and modulated inflammatory response in monocytes.^4,5 A Phase 2/3 randomized trial suggested possible efficacy in COVID-19 respiratory failure.⁶ Remdesivir, a small molecule antiviral, improves outcomes among hospitalized patients in general, but data among critically ill patients are less robust.⁷

ACTIV (Accelerating COVID-19 Therapeutic Interventions and Vaccines / Therapeutics for Inpatients with COVID-19) is an NIH-led public/private partnership established in April 2020 to advance COVID-19 therapeutics and vaccines. ACTIV-3b, Therapeutics for Severely Ill Inpatients with COVID-19 (TESICO), which focuses on patients with COVID-19 critical illness with hypoxemic respiratory failure, is one of several ACTIV therapeutic master protocols enabling simultaneous evaluation of multiple agents.⁸ The ACTIV agent selection committee selected aviptadil for study within TESICO.⁹ We report here the results of the trials comparing intravenous aviptadil and intravenous remdesivir each with placebo among patients with COVID-19 acute hypoxemic respiratory failure.

METHODS

Trial Design and Oversight

For the TESICO trial of aviptadil and remdesivir, depending on eligibility, consenting participants could be randomized to both study treatments or matching placebo in a 2×2 factorial design or to just one of the agents (details in the Supplementary Appendix, Section 2 and Figure S1). A factorial design was used for sample size efficiency because an interaction was considered unlikely. Randomization to only one of the two factors was permitted in order to include participants who were not eligible for one of the investigational agents.

The TESICO protocol was approved by a governing institutional review board for each enrolling site. Written informed consent for trial participation was obtained from each enrolled participant or a legally authorized representative. The trial was overseen by an independent data and safety monitoring board (DSMB) and conducted under an investigational new drug protocol with the U.S. Food and Drug Administration.

Target Population

Hospitalized adult patients were eligible if they had acute hypoxemic respiratory failure due to confirmed SARS-CoV-2 infection and were within four days of onset of respiratory failure. Acute hypoxemic respiratory failure was defined pragmatically as hypoxemia (defined clinically, with no specific threshold required) caused by COVID-19 pneumonia plus receipt of high-flow nasal oxygen (HFNO), non-invasive ventilation (NIV), invasive mechanical ventilation (IMV), or extracorporeal membrane oxygenation (ECMO).^10,11

Agent-specific exclusion criteria for aviptadil included refractory hypotension, severe diarrhea, or end-stage liver disease. Participants were excluded from the factorial randomization if they did not meet the eligibility criteria for both aviptadil and remdesivir. If eligible for one of the treatments, participants could be randomized separately to that active agent/placebo. A full list of eligibility criteria is provided in the Supplementary Appendix (Section 2).

Randomization and Blinding

Participants were randomized with a web-based application. For each site, randomization was stratified by disease severity (HFNO/NIV or IMV/ECMO), and four strata defined by remdesivir and aviptadil eligibility: 1) eligible for randomization to aviptadil and remdesivir in the 2×2 factorial; participants were equally randomized to either intravenous aviptadil, aviptadil matching placebo, intravenous remdesivir, or remdesivir matching placebo (1:1:1:1), with participants in the factorial receiving blinded infusions of both agents; 2) eligible for randomization to aviptadil only because remdesivir was started prior to randomization; 3) eligible for randomization to aviptadil only because remdesivir was contraindicated; and 4) eligible for randomization to remdesivir only because aviptadil was contraindicated (Figure S1). For participants in strata 2 through 4, randomization was (1:1) to the active agent or matching placebo. Permuted block randomization was used to generate the randomization schedules for each site (see Supplementary Appendix Section 2). Research staff, clinical personnel, and participants were blinded to trial group assignment; an unblinded pharmacist prepared blinded study product for infusion on site.

Interventions/Treatments

Aviptadil acetate was administered as a daily 12-hour infusion for three days, targeting 600 pmol/kg on infusion day 1, 1200 pmol/kg on infusion day 2, and 1800 pmol/kg on infusion day 3. Remdesivir was administered as a 200 mg loading dose, followed by 100 mg daily maintenance doses for up to a 10-day total course. For participants assigned placebo for either agent, matched saline placebo was administered in identical volumes.

All infusions of aviptadil/placebo were administered in an intensive care or step-down unit; participants discharged from intensive care or step-down unit did not continue dosing of aviptadil/placebo. If respiratory failure had resolved on or before Day 5, clinical teams were allowed to stop blinded remdesivir/placebo after 5 doses. Neither agent was continued after hospital discharge if the treatment course had not been completed.

Expected side effects of aviptadil included hypotension and diarrhea. Aviptadil/matching placebo was contraindicated on a given infusion day for participants receiving ≥ 0.1 mcg/kg/min of norepinephrine or equivalent vasopressor or with > 3 liquid stools in the preceding 24 hours. Blood pressure was monitored per local clinical guidelines but at least every 2 hours for the duration of the aviptadil infusion and for 2 hours thereafter. Study management guidelines stipulated responses to worsening hypotension with pauses, decreases, or discontinuation of the aviptadil/placebo infusion and to substantial diarrhea with oral loperamide and/or changes in the infusion. The grading system used for hypotension and other adverse events is given in the Supplemental Appendix (Section 2).

Sites were strongly encouraged to administer glucocorticoids per NIH treatment guidelines.¹² By design, participants were either receiving remdesivir prior to randomization or were randomized to remdesivir versus remdesivir placebo, unless contraindicated. Use of antiplatelet/anticoagulant treatment and immunomodulators were at the discretion of treating clinicians.

Primary and Secondary Outcomes for Each Treatment Comparison

The primary outcome was a 6-category ordinal outcome defining the participant’s status at Day 90: (1) at home (defined as the type of residence before hospitalization) and off oxygen (recovered) for at least 77 days, (2) at home and off oxygen for 49–76 days, (3) at home and off oxygen for 1–48 days, (4) not hospitalized but either on supplemental oxygen or not at home, (5) hospitalized or in hospice care, or (6) dead (see Supplementary Appendix Section 2). Mortality through Day 90 was a secondary endpoint; all participants were also followed for mortality through Day 180. Adverse events of any grade severity were collected during each infusion and 2 hours after infusion completion. Composite safety outcomes were assessed through Day 5, Day 28, and Day 90. These outcomes are defined in the Supplementary Appendix (Section 2).

Baseline Serologic and Virologic Assays

Serostatus (Genscript cPass and Quanterix HD-X) and antigen (Quanterix Simoa) testing were performed centrally on cryopreserved plasma obtained at baseline, while viral strain was determined using strain-specific PCR for Delta or Omicron variants on mid-turbinate samples obtained at baseline. Further details of laboratory methods are presented in the Supplementary Appendix (Section 2).

Statistical Analysis

The enrollment target was 640 participants for both the aviptadil versus matched placebo comparison and for the remdesivir versus matched placebo comparison. With the planned factorial design, it was assumed that the total sample size for the two trials would be substantially less than 1,280. This sample size target for each trial provided 80% power to detect a common odds ratio (OR) of 1.5 for the primary ordinal outcome at the (two-sided) 0.05 level of significance. More detailed sample size assumptions are provided in the Supplementary Appendix (Section 2).

The independent DSMB regularly reviewed interim analyses; on May 25, 2022 the DSMB reviewed a pre-specified futility analysis using conditional power estimates which were based on 70% of the planned number of participants in the aviptadil/placebo group with the Day 90 primary outcome. Following this review, the DSMB recommended stopping randomization of aviptadil versus placebo because it was highly unlikely that statistical significance would be achieved with full enrollment (conditional power was 12.4%, assuming an OR of 1.5 for the as-yet unobserved data). Randomization to the aviptadil trial of the master protocol was stopped at the time of this recommendation. At this review, the DSMB also noted that while there were no safety concerns for the remdesivir/placebo comparison, given the slow enrollment there may be operational reasons to close that trial as well. The investigator and sponsor agreed to end the remdesivir/placebo trial on June 9, 2022. Guidelines given to the DSMB for the futility assessment by the blinded investigator team are provided in the Supplementary Appendix (Section 2).

The analysis cohort for treatment comparisons of aviptadil versus matching placebo was randomized participants who received any amount of aviptadil or placebo (modified intention-to-treat). The primary analysis used a proportional odds model to estimate the OR of a better outcome on aviptadil compared to placebo for the 6-category ordinal outcome at Day 90.¹³ The model was stratified by disease severity (HFNO/NIV or IMV/ECMO) at study entry. Treatment comparisons for the primary outcome were based on participants for whom the Day 90 outcome was ascertained. An analysis that considered missing Day 90 outcome data was also carried out and is described in the Supplementary Appendix (Section 2). Death through Day 90 and through Day 180 and composite safety outcomes through Days 28 and 90 were compared between treatment groups using Cox proportional hazards models stratified by disease severity; hazard ratios (HRs) and 95% confidence intervals (CIs) are cited. Follow-up was censored using the date of withdrawal/loss-to-follow-up, or Day 90 (or 180), whichever was earlier. Treatment groups were also compared using Fine-Gray models stratified by disease severity at study entry to estimate sub-hazard ratios (sHR) for secondary outcomes that have been used in other trials, including time to discharge, time to discharge home, and time to discharge home for 14 days; cumulative incidence was estimated using the Aalen and Johansen method and compared using Gray’s method.^14–17 Twenty-three exploratory subgroup analyses (15 of which were prespecified) were performed to assess treatment effect heterogeneity by baseline characteristics. The heterogeneity of ORs and HRs for subgroups were assessed by including an interaction term in the corresponding regression models. One of the subgroups considered for the aviptadil comparison was defined according to whether remdesivir was randomized as part of the 2×2 factorial, whether remdesivir was initiated prior to randomization, or whether remdesivir was contraindicated. For participants enrolled in the 2×2 factorial, the effect of aviptadil versus matched placebo was also compared for those randomly assigned remdesivir with those assigned matching placebo for remdesivir. Results of the subgroup analyses should be interpreted with caution because there was no adjustment for type 1 error.

Analyses for remdesivir versus matching placebo were carried out with similar methods.

For both the aviptadil and remdesivir efficacy comparisons, an OR >1 for the primary endpoint and a sHR >1 for Fine-Gray models indicate superiority of the active treatment compared to placebo. Treatment comparisons for safety outcomes that include death are presented such that HRs <1 or ORs <1 indicate a more favorable outcome for the active treatment. Statistical analyses were performed using SAS (version 9.4) or R (version 4.1).

Role of the funding source

Investigators from NIH were directly involved in all aspects of this study, including study design, data collection, data analysis, data interpretation, and writing of the report. All analyses of biological material were done in a blinded manner at laboratories affiliated with the funding source; data were sent to the statistical and data management center at the University of Minnesota for linkage to the trial database. Several representatives from NIH are part of the writing group for the manuscript. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing official policies, either expressed or implied, of the National Institutes of Health or Department of Veterans Affairs.

RESULTS

Participants

From April 21, 2021 to May 24, 2022, participants were enrolled at 28 sites in the United States (Table S1). For the aviptadil comparison, 471 participants were randomized to aviptadil or matched placebo. The mITT population comprised 461 participants (Figures S1 and S2) who received at least a partial infusion of aviptadil (n=231) or aviptadil matched placebo (n=230).

For the remdesivir comparison, 87 participants (Figures S1 and S3) were randomized to remdesivir or matched placebo and all received some infusion of remdesivir (n=44) or remdesivir matched placebo (n=43). Eighty-five participants, i.e., those enrolled in the 2×2 factorial, were included in the mITT analyses for both agents. Reasons for not receiving any infusion of blinded aviptadil are presented in Table S2.

Baseline characteristics for the aviptadil/placebo comparison by treatment group are summarized in Table 1 and Tables S3–S11 for the mITT population. Overall, for the 461 participants in the mITT population, median age was 57 (IQR 46, 66) years; 39% (n=178) were female; 53% (n=246) reported non-White race or Hispanic ethnicity. Ninety-four percent (n=431) were in an intensive care unit at baseline, with 59% (n=271) receiving HFNO or NIV, 40% (n=185) receiving IMV, and 1% (n=5) ECMO. Median times from hospital admission and from onset of respiratory failure to randomization were 2 (IQR: 2,4) days and 2 (IQR: 2, 3) days, respectively, 39% (n=179) met Berlin criteria for ARDS, while 97% (n=445) of participants met modified Berlin criteria (bilateral infiltrates and SF ratio <315).¹¹ Ninety-five percent (n=440) of participants were prescribed glucocorticoids, and 95% (n=436) were prescribed antiplatelet/anticoagulant treatment at baseline. Most participants (74%, n=305 of 414 with positive nucleocapsid RT-PCR results) were infected with the Delta variant.

Table 1.

Baseline Characteristics: Modified Intention to Treat Cohort for Aviptadil versus Placebo Comparison.

Characteristic	Aviptadil n=231	Aviptadil Placebo n=230	Total N=461
Agent randomization/eligibility stratum, n(%)  Factorial (both agents)	39 (16.9)	46 (20.0)	85 (18.4)
Aviptadil only, remdesivir contraindicated	12 (5.2)	10 (4.3)	22 (4.8)
Aviptadil only, current/prior remdesivir use	180 (77.9)	174 (75.7)	354 (76.8)
Age in years, median (IQR)	58 (46, 67)	57 (46, 66)	57 (46, 66)
Female sex, n (%)	94 (40.7)	84 (36.5)	178 (38.6)
Race/ethnicity, n (%)  White	102 (44.2)	113 (49.1)	215 (46.6)
Black	40 (17.3)	33 (14.3)	73 (15.8)
Hispanic	62 (26.8)	57 (24.8)	119 (25.8)
Asian	7 (3.0)	10 (4.3)	17 (3.7)
Other	20 (8.7)	17 (7.4)	37 (8.0)
Body mass index in kg/m2, n (%)  30–39.9	92 (40.4)	89 (39.0)	181 (39.7)
≥40.0	55 (24.1)	61 (26.8)	116 (25.4)
Days since onset of respiratory failure, median (IQR)	2 (2, 3)	2 (2, 3)	2 (2, 3)
Baseline respiratory support, n (%)  HFNO device	127 (55.0)	118 (51.3)	245 (53.1)
NIV	9 (3.9)	17 (7.4)	26 (5.6)
IMV	93 (40.3)	92 (40.0)	185 (40.1)
ECMO	2 (0.9)	3 (1.3)	5 (1.1)
SF ratio (SPO2/FiO2) at randomization, median (IQR)	134 (100, 176)	131 (98, 186)	133 (99, 182)
Estimated PF ratio (PaO2/FiO2) at randomizationa, median (IQR)	83 (43, 133)	79 (40, 145)	74 (64, 86)
Bilateral lung infiltrates and SF ratio < 315, n (%)	224 (97.4)	221 (96.1)	445 (96.7)
ARDS by Berlin criteriab, n (%)	89 (38.7)	90 (39.1)	179 (38.9)
Vasopressor use at randomization, n (%)	35 (15.2)	29 (12.6)	64 (13.9)
Co-existing chronic illnessc n (%)  Hypertension	99 (42.9)	92 (40.0)	191 (41.4)
Diabetes mellitus	82 (35.5)	71 (30.9)	153 (33.2)
Renal impairment	45 (19.5)	38 (16.5)	83 (18.0)
Heart failure	24 (10.4)	15 (6.5)	39 (8.5)
Immunocompromisedd n (%)	36 (15.6)	32 (13.9)	68 (14.8)
SARS-CoV-2 Vaccination Statuse, n (%)  No vaccine dose received Medication use prior to randomization, n%	149 (64.5)	151 (65.7)	300 (65.1)
Remdesivir	177 (76.6)	172 (74.8)	349 (75.7)
Corticosteroid	219 (94.8)	221 (96.1)	440 (95.4)
Immunomodulator	78 (33.8)	78 (33.9)	157 (33.9)
Infecting variant (among the N=414 with positive nucleocapsid RT-PCRf), n (%)  Delta	155 (74.9)	150 (72.5)	305 (73.7)
Omicron	25 (12.1)	25 (12.1)	50 (12.1)
Other	27 (13.0)	32 (15.5)	59 (14.3)
Genscript neutralising anti-spike antibody positiveg, n (%)	157 (70.7)	156 (70.0)	313 (70.3)
BioRad anti-nucleocapsid antibody positiveh, n (%)	171 (77.0)	182 (81.6)	355 (79.3)
Nucleocapsid antigen concentration (pg/mL)I, median (IQR)	1246 (95, 6114)	1502 (224, 6406)	1294 (150, 6200)
Positive (concentration ≥3 pg/mL), n (%)	211 (95.0)	211 (94.6)	422 (94.8)

Abbreviations:HFNO, high flow nasal oxygen; NIV, non-invasive ventilation; IMV, invasive mechanical ventilation; ECMO; extracorporeal membrane oxygenation; IQR, interquartile range; n, number; kg, kilogram; m, meter; Covid-19, coronavirus disease 2019; RT-PCR, reverse transcription-polymerase chain reaction

^aEstimated PF ratio computed as PF ratio = (SF ratio – 64)/0.84. Rice TW, Wheeler AP, Bernard GR, et al. Comparison of the SpO2/FIO2 ratio and the PaO2/FIO2 ratio in patients with acute lung injury or ARDS. Chest. 2007 Aug;132(2):410–7. doi: 10.1378/chest.07-0617. Epub 2007 Jun 15. PMID: 17573487.

^bBilateral lung infiltrates and PF ratio < 300 and (IMV or ECMO) at baseline.

^cFull list of co-existing chronic illness in the Supplementary Appendix (Table S6)

^dImmunocompromised defined as receiving antirejection medications, biologic medications to treat autoimmune disease or cancer, human immunodeficiency virus, or other immunosupmpressive condition

^eNot vaccinated = No known vaccine doses received before randomization (includes 18 with unknown vaccination status; 6 aviptadil, 12 placebo)

^fSARS-CoV-2 Delta variant was determined from a mid-turbinate swab at baseline based on RT-PCR detection of the N-terminal domain of the Delta spike. Omicron variant was determined by similar method. Data available for 414; 207 aviptadil and 207 placebo.

^gGenScript cPass surrogate SARS-CoV-2 neutralization assay (anti-spike); positive: ≥30% binding inhibition. Data available for 445; 222 aviptadil, 223 placebo.

^hBioRad Platelia anti-nucleocapsid assay (total antibody); positive: ≥1.0 sample/cutoff ratio. Data available for 445; 222 aviptadil, 223 placebo.

ⁱQuanterix Simoa nucleocapsid antigen; positive: ≥3 pg/mL. Data available for 445; 222 aviptadil, 223 placebo.

Seventy-six percent (n=349) of participants assigned aviptadil or placebo were receiving remdesivir prior to randomization.

Baseline characteristics for the remdesivir/placebo comparison by treatment group are summarized in Section 5 of the Supplemental Appendix (Tables S12–S20).

Aviptadil

Adherence to Infusions and Concomitant Treatments

Infusions of aviptadil or matching placebo were closely monitored (Table S21). On the first infusion day (600 pmol/kg), 91% (n=211 aviptadil and n=210 placebo) participants received a full dose (defined as at least 90% of the prescribed volume), with no difference by treatment group. On subsequent infusion days (1200–1800 pmol/kg), fewer participants received a full dose (83% (n=190) for aviptadil vs. 91% (n=208) for placebo on the second day, 75% (n=170) vs. 88% (n=197) on the third day), with participants in the aviptadil group consistently receiving less study infusion than the placebo group.

The great majority of participants in both treatment groups continued glucocorticoids through Day 5 and antiplatelet/anticoagulant treatment through Day 7 (Table S22).

Efficacy Summary

The OR for being in a better category of the primary efficacy endpoint for aviptadil vs. placebo at Day 90, from a model stratified by baseline disease severity, was 1.11 (95% CI: 0.80–1.55; p=0.54) (Figure 1). The p-value corresponding to the test of proportional odds was 0.078; ORs for dichotomized outcomes for the 6 categories of the primary ordinal outcome are summarized in Table 2. The cumulative distribution of supplemental oxygen-free days at home through Day 90 (“days recovered”) is shown in Figure S4. The unadjusted OR for the 6-category Day 90 outcome, and the ORs after adjustment for other randomization stratification variables were consistent with the primary analysis (Table S23).

Figure 1.

Primary 6-category ordinal outcome at Day 90 for the aviptadil versus placebo comparison. Category percentages at Day 90 for the aviptadil group and the placebo group are shown. The summary odds ratio (OR) was estimated with the use of a proportional odds model that was stratified by disease severity at entry. An OR > 1.0 favors aviptadil. Estimated for the 450 participants with known status on Day 90. See Supplementary Appendix Figure S4 for related figure that includes imputation for those with unknown Day 90 status.

Table 2:

Summary of Major Efficacy Outcomes for the Aviptadil versus Placebo Comparison: Modified Intention to Treat Cohort

	Aviptadil		Placebo
Primary Outcome at Day 90	No. in Group	No. with Event (%in Group)	No. in Group	No. with Event (% in Group)	Odds Ratio for Aviptadil/Placebo (95% CI)	p value
6-category primary ordinal outcome at Day 90a,b	225		225		1.11 (0.80, 1.55)	0.54
Dichotomized outcomes for the 6-category primary ordinal outcome at Day 90a,c  Category 1  Category 2–6 (reference)  Category 1–2  Category 3–6 (reference)  Category 1–3  Category 4–6 (reference)  Category 1–4  Category 5–6 (reference)  Category 1–5  Category 6 (reference)	225	36 (16.0) 189 (84.0) 73 (32.4) 152 (67.6) 116 (51.6) 109 (48.4) 131 (58.2) 94 (41.8) 139 (61.8) 86 (38.2)	225	31 (13.8) 194 (86.2) 66 (29.3) 159 (70.7) 98 (43.6) 127 (56.4) 129 (57.3) 96 (42.7) 142 (63.1) 83 (36.9)	1.19 (0.71, 2.01) 1.16 (0.77, 1.73) 1.38 (0.95, 2.00) 1.04 (0.71, 1.51) 0.94 (0.64, 1.38)	0.51 0.48 0.09 0.85 0.77
Dichotomized outcomes for the 6-category primary ordinal outcome at Day 90a,c  Category 1  Category 2–6 (reference)  Category 1–2  Category 3–6 (reference)  Category 1–3  Category 4–6 (reference)  Category 1–4  Category 5–6 (reference)  Category 1–5  Category 6 (reference)	225	36 (16.0) 189 (84.0) 73 (32.4) 152 (67.6) 116 (51.6) 109 (48.4) 131 (58.2) 94 (41.8) 139 (61.8) 86 (38.2)	225	31 (13.8) 194 (86.2) 66 (29.3) 159 (70.7) 98 (43.6) 127 (56.4) 129 (57.3) 96 (42.7) 142 (63.1) 83 (36.9)	1.19 (0.71, 2.01) 1.16 (0.77, 1.73) 1.38 (0.95, 2.00) 1.04 (0.71, 1.51) 0.94 (0.64, 1.38)	0.51 0.48 0.09 0.85 0.77
Other Efficacy Outcomes Through Day 90	No. in Group	No. with Event (Estimated Cumulative % in Group)	No. in Group	No. with Event (Estimated Cumulative % in Group)	Hazard Ratio or Sub-Hazard Ratio for Aviptadil/Placebo (95% CI)	p value
Deathd	231	86 (37.5)	230	83 (36.2)	1.04 (0.77, 1.41)	0.78
Dischargede	231	139 (60.4)	230	138 (60.3)	0.99 (0.78, 1.24)	0.90
Discharged homee	231	132 (57.4)	230	133 (58.1)	0.97 (0.77, 1.23)	0.81
Discharged home for 14 consecutive days (sustained recovery) e	231	126 (55.1)	230	128 (56.2)	0.97 (0.76, 1.23)	0.78
Death, end-organ failure or serious infectiond	231	165 (71.6)	230	158 (69.0)	1.13 (0.91, 1.40)	0.29
Worsening respiratory failure or deathd	231	106 (46.2)	230	106 (46.3)	1.00 (0.77, 1.31)	0.98
Hospital readmission or death, after initial discharged,f	139	10 (7.8)	138	15 (11.2)	0.66 (0.30, 1.47)	0.31
Through Day 180
Deathd	231	90 (39.3)	230	86 (37.6)	1.06 (0.79, 1.42)	0.71

^aCategory 1: At home and off oxygen ≥ 77 days (best) Category 2: At home and off oxygen 49–76 days Category 3: At home and off oxygen 1–48 days Category 4: Discharged, but not at home, or at home requiring supplemental oxygen Category 5: Hospitalized or receiving hospice care Category 6: Died (worst)

^bSummary odds ratio for being in a better category, aviptadil vs. placebo. Proportional odds regression model with 1 indicator for treatment group stratified by disease severity. Computed for the 450 participants with known status at day 90; refer to the Supplementary Appendix (Figure S4) for the odds ratio after imputation for those with missing data.

^cOdds ratio for being in a better category compared to designated reference group, aviptadil vs. placebo. Logistic regression model with 1 indicator for treatment group stratified by disease severity.

^dHazard ratio for time to first event, aviptadil vs. placebo. Cox proportional hazards regression model with 1 indicator for treatment group stratified by disease severity.

^eSub-hazard ratio for time to first event, aviptadil vs. placebo. Fine-Gray model considering death a competing risk with 1 indicator for treatment group stratified by disease severity.

^fAmong participants who were discharged from the index hospital. Time=0 set to the date of discharge.

Eleven participants (2% of the mITT population) with follow-up through Day 90 (6 assigned aviptadil and 5 placebo) had an unknown status for the Day 90 ordinal primary outcome. Sensitivity analyses that imputed the outcome category for participants with unknown Day 90 status were consistent with the primary analysis (OR=1.10; 95% CI: 0.79–1.52; Figure S5).

Through Day 90, 86 participants in the aviptadil group and 83 in the placebo group died. The cumulative percent who died through Day 90 was 38% and 36% in the aviptadil and placebo groups respectively (Figure 2) with a HR for aviptadil compared to placebo of 1.04 (95% CI: 0.77–1.41). By Day 180, the percentages who died were 39% (90 deaths) and 38% (86 deaths), respectively with a corresponding HR of 1.06 (95% CI: 0.79, 1.42).

Figure 2.

Kaplan-Meier estimates of cumulative mortality through Day 180 for the aviptadil group and the placebo group. Hazards ratio (HR) estimated with a proportional hazards regression model stratified by disease severity at entry. A HR > 1.0 favors placebo. Estimated for all 461 in the mITT cohort.

Secondary efficacy outcomes are further summarized in Table 2, Figures S6–S13 and Table S24. For each of these measures, the 95% CI for the relative treatment difference includes 1.0.

Safety Summary

Infusion reactions were significantly more common for participants treated with aviptadil compared to placebo (Table 3 and Tables S25–S32). Twenty-four (10%) aviptadil and 13 (6%) placebo participants had an infusion discontinued due to an adverse event; 71 (31%) and 35 (15%) experienced an infusion pause due to an adverse event (Table 3). Most adverse events during or within 2 hours of completing the infusion were grade 1 or 2. Diarrhea, facial flushing, tachycardia and hypotension occurred more frequently on aviptadil than placebo; these events were more common for aviptadil than placebo on the second and third infusion days, when higher aviptadil doses were targeted (Table 3).

Table 3.

Safety Outcomes for the Aviptadil versus Placebo Comparison: Modified Intention to Treat Cohort

Outcomes	Aviptadil (n=231), n (%)	Placebo (n=230), n (%)	Odds or Hazard Ratio for Aviptadil/Placebo (95% CI)	p value
Infusion reactions of any grade, Days 0–2 a
Any reaction	200 (86.6)	132 (57.4)	--	<0.001
Facial flushing	34 (14.7)	14 (6.1)	--	0.002
Diarrhea	92 (39.8)	26 (11.3)	--	<0.001
Tachycardia	22 (9.5)	11 (4.8)	--	0.049
Hypotension	135 (58.4)	95 (41.3)	--	<0.001
Infusion adjustments due to AEb, Days 0–2
Pause	71 (30.7)	35 (15.2)	--
Reduction in rate	53 (22.9)	20 (8.7)	--
Discontinuation	24 (10.4)	13 (5.7)
Safety through Day 5 c
Composite safety outcome of SAE, Grade 3/4 AE, Organ Failure/Serious Infection, or Death	146 (63.2)	129 (56.1)	1.	0.10
Select components
Deathc	13 (5.6)	11 (4.8)	1.19 (0.52, 2.71)	0.68
Hypotension or shockc,d	69 (29.9)	64 (27.8)		0.62
Diarrheac,e	4 (1.7)	2 (0.9)		0.42
Cardiacc,f	24 (10.4)	22 (9.6)	--	0.76
Renalc,g	20 (9.4)	12 (5.4)	--	0.10
Barotraumac,h	5 (2.2)	1 (0.4)		0.14
Safety through Day 28 i
Composite safety outcome of SAE, Grade 3/4 AE, Organ Failure/Serious Infection, or Death	181 (78.4)	172 (74.8)	1.17 (0.95,1.44)	0.15
Select components
Deathi	69 (29.9)	66 (28.7)	1.05 (0.75, 1.47)	0.77
Hypotension or shockd,i	112 (48.5)	114 (49.6)	--	0.82
Diarrheae,i	7 (3.0)	4 (1.7)	--	0.37
Cardiac f,i	56 (24.2)	49 (21.3)	--	0.49
Renalg,i	41 (19.2)	40 (17.9)		0.56
Barotrauma h,i	11 (4.8)	14 (6.1)		0.55

SAE=serious adverse event; AE=adverse event; MedDRA=Medical Dictionary for Regulatory Activities; SOC=System Organ Class.

^aP-value for comparing infusion reactions between treatment groups from Cochran-Mantel-Haenszel tests, stratified by disease severity at baseline.

^bInfusion adjustments are not mutually exclusive, see Table S17 for mutually exclusive representation.

^cOdds ratio for experiencing the event, aviptadil vs. placebo. Logistic regression model with 1 indicator for treatment group, stratified by disease severity at baseline. P-values for treatment group comparisons for select components from separate similar logistic regression models.

^dAny grade 3/4 hypotension during infusion, any organ failure report of hypotension with vasopressor use, any grade 3/4 AE, SAE, or death with MedDRA PT for hypotension, shock, shock haemorrhagic, distributive shock

^eAny grade 3/4 AE report if diarrhea. Per protocol, infusion-related diarrhea events excluded unless event led to discontinuation of blinded aviptadil.

^fAny grade 3/4 AE, SAE, or death with MedDRA Cardiac SOC, and any organ failure report of myocardial infarction, congestive heart failure III/IV, or atrial/ventricular tachyarrhythmia

^gAny grade 3/4 AE, SAE, or death with MedDRA PT for acute kidney injury, and organ failure report of new renal replacement therapy. Limited to patients without dialysis at baseline, Aviptadil= 213 and Placebo= 223

^hAny grade 3/4 AE, SAE, or death with MedDRA PT for pneumomediastinum or pneumothorax.

ⁱHazard ratio for time to first event, aviptadil vs. placebo. Cox proportional hazards regression model with 1 indicator for treatment group stratified by disease severity. P-values for treatment group comparisons for select components from separate similar logistic regression models.

On average, on the second and third infusion days, mean arterial pressure (MAP) was 5–7mmHg lower among aviptadil vs. placebo participants during each infusion (p<0.001), but returned to parity by 2 hours after completing the infusion (Figure 3). Further details on hypotension adverse events associated with the infusion are provided in Table S32.

Figure 3.

Average change in mean arterial pressure (MAP) and systolic blood pressure (SBP) during the blinded aviptadil infusions on days 0, 1, and 2. MAP and SBP were reported pre-infusion (hour=0), every 2 hours peri-infusion, and 2 hours after the infusion ended.

The primary safety outcome of death, serious adverse events, organ failure, serious infection, or grade 3 or 4 adverse events through Day 5 (Tables 3 and S33–S46) was experienced by 63% (n=146) of participants in the aviptadil compared to 56% (n=129) in the placebo groups, OR 1.40 (95% CI 0.94–2.08), p=0.10. Using data through Day 28, the percentages were 78% (n=181) versus 75% (n=172) in the aviptadil and placebo groups, respectively, with HR 1.17 (95% CI 0.95–1.44), p=0.15 (Tables 3 and S47–S61 and Figure S14). While hypotension was more common in the aviptadil than placebo groups during the infusion, the percentage of participants with potentially serious renal complications of hypotension through Day 28, among those without dialysis at entry, were similar for the aviptadil and placebo groups, 19% (n=41/213) and 18% (n=40/223), respectively (Table 3).

Summaries of serious adverse events, including organ failure and serious infections, and cardiovascular events through Day 90, and laboratory abnormalities through Day 5 are in the Supplemental Appendix (Tables S62–S66).

Subgroup Analyses

The estimated OR (aviptadil/placebo) for the primary ordinal outcome did not vary according to disease severity (Figure S15). ORs for those receiving HFNO or NIV at entry and IMV or ECMO at entry were 1.04 (95% CI: 0.67–1.60) and 1.21 (95% CI: 0.73–2.04), respectively (p=0.71 for interaction). ORs for the primary endpoint varied across four subgroups defined by use (or nonuse) of remdesivir (p=0.038). Among the 84 participants enrolled in the factorial, ORs (aviptadil/placebo) were 2.11 (95% CI: 0.69–6.46) for those randomized to remdesivir and 3.78 (95% CI: 1.17–12.24) for those randomized to the remdesivir placebo. For those for whom remdesivir was contraindicated (n=22), the OR was 0.64 (95% CI: 0.10–4.08); and for those from whom remdesivir had been started prior to randomization (n=344), the majority of participants, the OR was 0.93 (95% CI: 0.64–1.36). For the 84 participants in the factorial study, ORs (aviptadil/placebo) for those also assigned remdesivir as compared to placebo for remdesivir did not vary (p=0.33) (Figure S15).

There was also possible evidence of heterogeneity for subgroups defined by race/ethnicity. This apparent heterogeneity arose because of an OR (2.71, 95% CI: 1.37–5.33) indicating a possible favorable effect of aviptadil for those of non-Black Hispanic ethnicity and ORs indicating possible unfavorable effects (OR 0.78, 95% CI: 0.34–1.83 for Black participants, and OR 0.76, 95% CI: 0.46–1.23 for non-Hispanic White participants) for other race/ethnicity groups (Figure S15). Subgroup analyses for Day 90 mortality (Figure S16 and Table S67) and the Day 5 and 28 composite safety outcomes are summarized in Figures S17–S18.

Follow-up results for the 84 participants randomized to remdesivir vs. placebo are presented in Section 7 of the Supplementary Appendix.

DISCUSSION

In this multicenter phase III trial of participants hospitalized for COVID-19-associated respiratory failure, treatment with aviptadil did not improve the primary ordinal outcome or survival at Day 90 compared to placebo. Other secondary endpoints also did not demonstrate efficacy for aviptadil. Safety analyses were largely consistent with the known safety profile of aviptadil—hypotension, diarrhea, and facial flushing were the most common and prominent events associated with infusion.

This trial specifically addresses patients with critical COVID-19 acute hypoxemic respiratory failure, who suffer the highest COVID-19-related mortality.^18–20 In vitro, aviptadil has four potential mechanisms of action relevant to COVID-19 acute hypoxemic respiratory failure, including suppression of SARS-CoV-2 replication, decrease in viral cytopathy, modulation of monocyte-derived inflammation, and increases in surfactant production.⁵ Based on multiple complementary mechanisms of action, experience with pre-pandemic ARDS, and an exploratory efficacy signal from a previous Phase 2/3 randomized trial,⁶ aviptadil was a logical lung-specific agent for investigation in COVID-19 acute hypoxemic respiratory failure. While in the present study aviptadil infusion appeared safe in a closely monitored, largely ICU, setting in which on-treatment hypotension can be promptly addressed, we found no evidence to suggest that aviptadil has a therapeutic role for patients with COVID-19 acute hypoxemic respiratory failure.

Although more than half of enrolled participants reported Hispanic ethnicity or non-White race, the trial had limited power to explore subgroups based on race/ethnicity. A possible favorable effect of aviptadil observed among Hispanic participants (and a possibly similar effect among patients in the remdesivir factorial, who were more often Hispanic) must be interpreted with caution, due to the absence of an overall treatment effect as well as the large number of subgroups assessed.

Strengths and limitations

This trial demonstrates the feasibility of robust trials specific to critical COVID-19 acute hypoxemic respiratory failure, with enhanced safety monitoring adapted to the specific patient population. Study drug was infused with intensive monitoring, structured management guidance for hypotension, and specific grading tables for hypotension.

The safety protocols used within this trial aimed to balance potential risks and benefits, emphasizing infusion pauses and dose reduction rather than substantial increases in vasopressor therapy and/or fluid loading to counter the hypotensive effects of aviptadil, an investigational agent with unproven clinical benefit. This approach resulted in 24% of assigned participants not receiving the full prescribed dose of aviptadil across the three infusion days, although 97% of participants received 50% or more of the target dose. While it is therefore possible that the dose or duration of aviptadil used (3 days of escalating doses, 12 hours per day) was not ideal, we employed dosing conventions used in prior studies at the time of study launch that pharmacokinetic modeling suggests achieve adequate lung concentrations.^5,21,22 Furthermore, the average decrease in MAP of 5–7 mmHg on the second and third days of infusion suggests that a higher dose of aviptadil would not be well tolerated. Given the lack of efficacy and increasing intolerance on later dosing days, we believe it unlikely that longer courses of treatment would have a benefit.

Methodological innovations include use of a novel, patient-centered endpoint and a pragmatic definition of respiratory failure that included patients receiving HFNO.^10,11 The primary endpoint combined mortality and the speed with which survivors recovered; the definition of recovery was patient-centered given the substantial burden of new supplemental oxygen and/or failure to return home. The inclusion of patients receiving HFNO is justified based on the homogeneity of this population (all had COVID-19 pneumonia as the cause of their hypoxemic respiratory failure), the fact that 40% of patients receiving HFNO at baseline subsequently underwent invasive mechanical ventilation, and the concordance of results between patients receiving HFNO or invasive mechanical ventilation at baseline.

Additional strengths of this trial include rigorous safety protocols for the study of a vasodilating agent in patients with or at risk for hypotension, the racial/ethnic diversity of participants, use of evidence-based standard of care therapy, and standardized ascertainment of endpoints. Specifically, ascertainment of the primary, patient-centered endpoint through Day 90 was >95%, suggesting that (with appropriate resources and training) outcomes beyond hospital mortality can be reliably measured in this complex and critically ill population.

Our trial has other limitations. First, the trial has limited power for subgroup analyses. Second, while the trial was performed after SARS-CoV-2 vaccines were generally available, most participants had not been vaccinated, limiting generalization to a more vaccinated population. Third, while the vast majority of trial participants received corticosteroids, NIH COVID-19 treatment guidelines changed during the conduct of the TESICO trial, adding second-line immune-suppression for hospitalized patients with rapidly worsening hypoxemia, and only 34% (n=156) of trial participants received second-line immune-suppression at enrollment. Fourth, enrollment in the remdesivir factor of the 2×2 factorial was limited by extensive pretreatment in the patient population, and as a consequence, planned comparisons of remdesivir with placebo are substantially underpowered. Fifth, we did not collect formal screening logs from enrolling sites, although we enrolled with broad eligibility criteria at diverse sites, and the enrolled trial population was ethnically diverse, suggesting good generalizability.

In summary, we found no evidence that intravenous aviptadil improved clinical outcomes among patients with COVID-19-associated acute hypoxemic respiratory failure; inferences regarding remdesivir are limited by sample size.

Research in Context.

Evidence before this study

Mortality remains high among hospitalized patients with COVID-19 who experience critical hypoxemic respiratory failure. While certain immunomodulators (especially glucocorticoids and perhaps baricitinib and tocilizumab) have suggested efficacy in this critically ill population, mortality remains high. Lung-specific therapies are an important priority for patients with this condition, given the strong association between the severity of lung failure and ultimate patient outcomes. Aviptadil, the synthetic form of Vasoactive Intestinal Peptide, is a neuropeptide hormone with positive pleiotropic effects in preclinical models. A phase 2/3 trial of intravenous aviptadil among hospitalized patients with COVID-19 and respiratory failure suggested the possibility of efficacy for intravenous aviptadil in this population. In addition, prior trials suggested the possibility that the antiviral remdesivir may have efficacy in this population, although this is not certain. We serially searched MEDLINE in English from 1970 to 2021 for (“aviptadil” or “remdesivir”) and (“COVID” or “novel coronavirus” or SARS-CoV-2).

Added value of this study

In this multicenter Phase 3 trial, we evaluated three daily infusions of intravenous aviptadil vs. placebo and/or 10 days of intravenous remdesivir in the setting of standard care (including glucocorticoid therapy). Aviptadil did not improve the primary endpoint of recovery status at Day 90; nor did aviptadil improve any other endpoint. The anticipated safety profile was confirmed—flushing, hypotension, and diarrhea were all more common in the aviptadil-assigned group. Few patients were randomized to the remdesivir vs. placebo comparison.

Implications of all the available evidence

Aviptadil does not appear to have a role in the treatment of critical respiratory failure among patients hospitalized with COVID-19. Other lung-related therapies should be sought in this severely underserved population.

DATA SHARING STATEMENT

Deidentified data and supporting documentation, including the protocol, statistical analysis plan, informed consent document, and data dictionary will be made available to researchers after approval of a proposal for use of the data. Proposals for data use should be submitted by means of the research proposal form on the INSIGHT website (www.insight-trials.org).