Abatacept, Cenicriviroc, or Infliximab for Treatment of Adults Hospitalized With COVID-19 Pneumonia: A Randomized Clinical Trial

Key Points

Question  Do abatacept, cenicriviroc, or infliximab improve time to recovery for patients hospitalized with COVID-19 pneumonia compared with standard care?

Findings  A randomized, double-masked, placebo-controlled master protocol clinical trial found that treatment with abatacept, cenicriviroc, or infliximab showed no statistically significant difference for the primary end point of time to recovery in patients with COVID-19 pneumonia vs standard care.

Meaning  Abatacept, cenicriviroc, or infliximab did not decrease time to recovery for hospitalized patients with COVID-19 pneumonia.

Importance  Immune dysregulation contributes to poorer outcomes in COVID-19.

Objective  To investigate whether abatacept, cenicriviroc, or infliximab provides benefit when added to standard care for COVID-19 pneumonia.

Design, Setting, and Participants  Randomized, double-masked, placebo-controlled clinical trial using a master protocol to investigate immunomodulators added to standard care for treatment of participants hospitalized with COVID-19 pneumonia. The results of 3 substudies are reported from 95 hospitals at 85 clinical research sites in the US and Latin America. Hospitalized patients 18 years or older with confirmed SARS-CoV-2 infection within 14 days and evidence of pulmonary involvement underwent randomization between October 2020 and December 2021.

Interventions  Single infusion of abatacept (10 mg/kg; maximum dose, 1000 mg) or infliximab (5 mg/kg) or a 28-day oral course of cenicriviroc (300-mg loading dose followed by 150 mg twice per day).

Main Outcomes and Measures  The primary outcome was time to recovery by day 28 evaluated using an 8-point ordinal scale (higher scores indicate better health). Recovery was defined as the first day the participant scored at least 6 on the ordinal scale.

Results  Of the 1971 participants randomized across the 3 substudies, the mean (SD) age was 54.8 (14.6) years and 1218 (61.8%) were men. The primary end point of time to recovery from COVID-19 pneumonia was not significantly different for abatacept (recovery rate ratio [RRR], 1.12 [95% CI, 0.98-1.28]; P = .09), cenicriviroc (RRR, 1.01 [95% CI, 0.86-1.18]; P = .94), or infliximab (RRR, 1.12 [95% CI, 0.99-1.28]; P = .08) compared with placebo. All-cause 28-day mortality was 11.0% for abatacept vs 15.1% for placebo (odds ratio [OR], 0.62 [95% CI, 0.41-0.94]), 13.8% for cenicriviroc vs 11.9% for placebo (OR, 1.18 [95% CI 0.72-1.94]), and 10.1% for infliximab vs 14.5% for placebo (OR, 0.59 [95% CI, 0.39-0.90]). Safety outcomes were comparable between active treatment and placebo, including secondary infections, in all 3 substudies.

Conclusions and Relevance  Time to recovery from COVID-19 pneumonia among hospitalized participants was not significantly different for abatacept, cenicriviroc, or infliximab vs placebo.

Trial Registration  ClinicalTrials.gov Identifier: NCT04593940

The immunopathology of COVID-19 involves aberrant immune cell function and dysregulation of cytokine and chemokine networks.^1-3 The RECOVERY trial demonstrated a mortality benefit for dexamethasone for hospitalized participants receiving supplemental oxygen, making it a mainstay of anti-inflammatory treatment for COVID-19.⁴ Despite therapeutic advances, considerable morbidity and mortality remain, spurring continued evaluation of therapeutic approaches. Medications disrupting IL-6 signaling (tocilizumab) and the Janus kinase pathway (baricitinib) show benefit in the most severely ill patients with COVID-19.^5-9 In April 2020, the US National Institutes of Health (NIH) launched a public-private partnership, Accelerating COVID-19 Therapeutic Interventions and Vaccines (ACTIV), to develop a coordinated research response to COVID-19. The ACTIV-1 Immune Modulator (IM) master protocol evaluated immunomodulatory agents in hospitalized participants with moderate/severe COVID-19. Three agents were selected for study based on their novel mechanisms of action and in vitro data against SARS-CoV-2. These included abatacept (T-cell costimulatory modulator that mitigates T-cell responses), cenicriviroc (CCR2/CCR5 antagonist that reduces monocyte and macrophage functions while sparing neutrophil and T-cell function), and infliximab (tumor necrosis factor α [TNF] inhibitor that binds and neutralizes soluble and transmembrane TNF).^3,10 Considerable safety data exist for abatacept and infliximab for other inflammatory disorders. This study reports the results of 3 randomized, double-masked, placebo-controlled substudies of abatacept, cenicriviroc, and infliximab plus standard care for COVID-19 pneumonia.

ACTIV-1 IM is a master protocol designed to evaluate multiple immune modulators plus standard care in participants hospitalized with COVID-19 through randomized, double-masked comparisons with placebo.¹¹ Data from participants receiving placebo were shared among the comparisons to minimize the number of participants receiving placebo, thereby reducing the sample size needed for adequate power for each comparison. The protocol and statistical analysis plan are available in Supplement 1.

The protocol was approved by relevant review boards; all participants provided written informed consent. Participants 18 years or older with confirmed SARS-CoV-2 infection within 14 days, anticipated hospital stay of 72 hours or more, and evidence of pulmonary involvement were eligible. Candidates with pregnancy, liver enzymes more than 10 times the normal value, chronic liver disease, acute kidney injury with glomerular filtration rate less than 30 mL/min (stable chronic kidney insufficiency permitted), severe heart failure, severe neutropenia, lymphopenia, or known or suspected untreated infection including tuberculosis and those who received cytotoxic or biologic-targeted immunomodulators within 4 weeks or 5 half-lives before screening were excluded. Substudy-specific exclusion criteria are outlined in Supplement 1.

Demographics and clinical status were extracted from the medical records when available or verified with participants. This included race and ethnicity, which were collected to allow for assessment of study participant representativeness. Study drug or placebo was administered according to substudy randomization. Abatacept (10 mg/kg; maximum dose, 1000 mg) and infliximab (5 mg/kg) were given as a single infusion; cenicriviroc (300-mg loading dose followed by 150 mg twice daily) was an oral medication administered for 28 days (Supplement 1). Participants received standard care at site hospitals, including remdesivir (study provided [Gilead Sciences]) and corticosteroids. Clinical status and safety data were captured daily during hospitalization through day 28 and postdischarge days 8, 11, 15, 29, and 60.

Eligible participants were randomized in a 2-stage process (Figure 1). First, each participant was assigned with equal probability to receive 1 of the immunomodulatory agents (substudy) using an open-label design. Second, each participant was randomized in a masked fashion to the test agent or its matching placebo in an n:1 ratio, where n equals the number of agents for which the participant was eligible.

Data from participants in the placebo group were shared for analysis of each agent once agent-specific eligibility criteria were applied (Supplement 1). Several elements were key to preserving the statistical integrity of shared placebos.¹¹ To avoid introduction of bias related to evolution of disease or standard care over time, the shared placebo group for each substudy analysis only included participants enrolled at a site when that substudy was available at the time of enrollment. The shared placebo group used for substudy analysis was limited to participants who met enrollment criteria for that substudy; participants could be included in the placebo group for 1, 2, or 3 substudies. However, if, for example, the participant did not meet agent-specific criteria for infliximab, they were not included in the placebo group for the infliximab substudy, but could be included in the placebo group for other agents. As a result, the placebo groups for each substudy were similar but not identical. Further, we did not use precision approaches to participant selection, such as biomarkers, that might affect agent-specific randomization.

The primary outcome of median time to recovery by day 28 was evaluated using an 8-point ordinal scale (higher scores indicate better health; Supplement 2).⁷ Recovery was identified as the first day a participant attained a score of 6, 7, or 8. Key secondary end points were prespecified in the statistical analysis plan (Supplement 1) as clinical status at day 14, assessed by ordinal scale improvement from baseline, and 28-day mortality. Other prespecified secondary end points were 14-day mortality, 60-day mortality, and clinical status at day 28.

Sample size requirements were based on the ability to detect moderate improvement in time to recovery. For each substudy, an estimated 788 recoveries were required to provide approximately 85% power to detect a recovery rate ratio (RRR) of 1.25 for active study drug vs placebo.⁷ The primary efficacy analysis was based on the Fine-Gray model.¹² Efficacy monitoring used the Lan-DeMets¹³ alpha spending function analog to the O’Brien-Fleming¹⁴ boundary. Futility monitoring used the moderately aggressive, Hwang-Shih-DeCani beta spending function with gamma parameter 2. A gatekeeping approach required the boundary for the primary end point to be crossed to declare a statistically significant result for the 2 key secondary end points (eTables 2, 8, and 14 in Supplement 2). The statistical analysis plan describes the type I error protection for the primary and the 2 key secondary end points (28-day mortality and day 14 clinical status; Supplement 1). Day 28 mortality was analyzed as a binary end point with an indicator variable for treatment group using a logistic regression model. Clinical status at day 14 was analyzed using a proportional odds model on the 8-point ordinal scale. A multiple-imputation approach was used to account for uncertainty of outcomes for participants lost to follow-up (Supplement 1).

The intention-to-treat (ITT) population was used for primary end point analysis as specified in the statistical analysis plan (Supplement 1). All secondary end points and safety analyses were reported for the “treated-as-assigned” population consisting of all randomized participants who received at least 1 dose of the assigned study drug.^15,16 All P values presented are 2-sided. All ITT and treated-as-assigned data for the primary and secondary end points were consistent and are presented.

Safety assessments included a composite end point of death, serious adverse events (SAEs), or grade 3 (severe) and 4 (potentially life-threatening) AEs through day 60. Secondary infections were collected as AEs of special interest through day 60 and were adjudicated by an independent safety officer.

Between October 16, 2020, and December 31, 2021, a total of 1971 participants underwent randomization at 95 hospitals (85 clinical research sites) across the ACTIV-1 IM platform (Figure 1). Baseline characteristics for each substudy are presented in Table 1 (treated-as-assigned population is shown in eTable 1 in Supplement 2). Characteristics by region are presented in eTables 3, 4, 9, 10, 15, and 16 in Supplement 2. The abatacept and infliximab substudies completed enrollment when they met the prespecified number of recoveries. On September 2, 2021, at the second interim analysis, the data and safety monitoring board recommended discontinuation of randomization to the cenicriviroc substudy due to futility. Treatment and subsequent follow-up were continued for participants already enrolled in this substudy and receiving study medication. No unmasking of individual treatment assignments was deemed necessary for cenicriviroc.

Abatacept did not meet the primary end point of time to recovery (RRR, 1.12 [95% CI, 0.98-1.28]; P = .09) after achieving 811 recoveries in the ITT population (Table 2, Figure 2, and eFigure 1 in Supplement 2). Median time to recovery was the same for abatacept compared with placebo (9 days [95% CI, 8-10]). Across ordinal scale subgroups, the P value for interaction was .36 (Table 2; eTable 2, eFigure 1 in Supplement).

Clinical status was assessed by proportional odds of improvement at day 14 (odds ratio [OR], 1.19 [95% CI, 0.94-1.50]) (Table 2 and eFigure 2 in Supplement 2) and day 28 (OR, 1.35 [95% CI, 1.06-1.73]) (eTable 5 in Supplement 2) in the treated-as-assigned population.

All-cause 28-day mortality for participants receiving abatacept was 11.0% compared with 15.1% for those receiving placebo (OR, 0.62 [95% CI, 0.41-0.94]) (Table 2, Figure 2, and eFigures 3-6 in Supplement 2) in the treated-as-assigned population. All-cause 14-day mortality for abatacept was 4.7% compared with 7.8% for placebo (OR, 0.55 [95% CI, 0.32-0.96]), and all-cause 60-day mortality rates were 14.5% for abatacept vs 17.1% for placebo (OR, 0.74 [95% CI, 0.50-1.08]) (eTable 5 and eFigure 7 in Supplement 2).

There was no statistically significant difference in the composite safety end point for participants who received abatacept (33.2%) vs placebo (34.9%) (risk difference, −1.7% [95% CI, −7.5% to 4.1%]). SAEs and grade 3 or 4 AEs were not different between the 2 groups. Having any SAE was reported for 128 participants (25.1%) in the abatacept group and 136 (26.7%) in the placebo group. Nine participants (1.8%) in the abatacept group and 7 (1.4%) in the placebo group had an SAE related to the study drug; none were deemed by the sponsor to be suspected unexpected serious adverse reactions related to abatacept administration. Three serious infusion reactions were reported for abatacept and formally adjudicated. One grade 4 SAE met the case definition for anaphylaxis and 2 grade 3 SAEs met criteria for infusion reactions.¹⁷ Safety assessment results are presented in Table 3.

Secondary infections by day 60 were reported for 82 participants (16.1%) in the abatacept group and 73 (14.3%) in the placebo group, with the types of infections being similar to those in the infliximab substudy (Table 3 and eTable 6 in Supplement).

Overall, 98.4% of participants in the cenicriviroc substudy received a loading dose of the study drug or placebo, with 96.7% receiving at least 1 subsequent maintenance dose. The median (IQR) duration of treatment was 28 (13-29) days. A total of 36.5% of participants discontinued maintenance doses prior to day 29.

There was no difference in time to recovery for cenicriviroc compared with placebo (RRR, 1.01 [95% CI, 0.86-1.18]; P = .94) with a median of 8 days (95% CI, 8-9 vs 8-10) for the ITT population (Table 2, Figure 2; eFigure 8 in Supplement 2).

There was no difference in the odds of improvement in clinical status at day 14 or day 28 assessed by the ordinal scale (day 14: OR, 0.93 [95% CI, 0.71-1.23]; day 28: OR, 0.89 [95% CI, 0.66-1.19]) (Table 2 and eTable 11 and eFigure 10 in Supplement 2) in the treated-as-assigned population.

Mortality through day 28 for the treated-as-assigned population was 13.8% for cenicriviroc and 11.9% for placebo (OR, 1.18 [95% CI, 0.72-1.94]) (Table 2, Figure 2, and eFigure 9 in Supplement 2). Day 14 mortality rates were 8.7% for participants receiving cenicriviroc vs 7.1% for those receiving placebo (OR, 1.31 [95% CI, 0.71-2.40]) and day 60 mortality rates were 18.0% for cenicriviroc and 13.8% for placebo (OR, 1.39 [95% CI, 0.89-2.18]) (eTable 11 and eFigure 11 in Supplement 2).

There was no clinically significant difference in the composite safety end point at day 60 between the groups (incidence of 36.9% for cenicriviroc vs 33.9% for placebo; risk difference, 3.0% [95% CI, −4% to 10%]). Grade 3 or 4 AEs occurred with similar frequencies (105 [29.6%] in the cenicriviroc group and 97 [27.4%] in the placebo group) (Table 3). SAEs occurred in 102 participants (28.7%) in the cenicriviroc group and 84 (23.7%) in the placebo group. The percentage of participants who had any secondary infection was similar at day 60 for cenicriviroc and placebo (53 [14.9%] vs 52 [14.7%]) (Table 3 and eTable 12 in Supplement 2).

Infliximab did not meet the primary end point of time to recovery in the ITT population, with an estimated RRR for infliximab vs placebo of 1.12 (95% CI, 0.99-1.28; P = .08) after achieving 826 recoveries (Table 2; eFigure 12 in Supplement 2). Median time to recovery for infliximab was similar to placebo (median of 8 [95% CI, 7-9] vs 9 [95% CI, 8-10] days). Results by ordinal scale are shown in Table 2 and eFigure 12 in Supplement 2. Across ordinal scale subgroups, the P value for interaction was .36, indicating no clear evidence of heterogeneity (eTable 14 in Supplement 2).

Clinical status was assessed as a proportional odds of improvement at day 14 (OR, 1.32 [95% CI, 1.05-1.66]) (Table 2 and eFigure 14 in Supplement 2) and day 28 (OR, 1.45 [95% CI, 1.14–1.85]) (eTable 17 in Supplement 2) for the treated-as-assigned population.

Mortality through day 28 in participants receiving at least 1 dose of the study medication (treated-as-assigned population) was 10.1% for infliximab and 14.5% for placebo (OR, 0.59 [95% CI, 0.39-0.90]) (Table 2, Figure 2, and eFigures 13 and 15-17 in Supplement 2). Fourteen-day mortality rates were 5.6% for infliximab vs 8.1% for placebo (OR, 0.63 [95% CI, 0.36-1.08]) and 60-day mortality rates were 12.6% for infliximab compared with 16.5% for placebo (OR, 0.68 [95% CI, 0.46-1.00]) (eTable 19 and eFigure 18 in Supplement 2).

No difference was observed in the composite safety end point at day 60 (incidence, 32.9% vs 33.7%; risk difference, −0.8% [95% CI, −6.6% to 4.9%]). SAEs occurred in 125 participants (24.2%) in the infliximab group, with 7 events in 6 participants (1.2%) determined by investigators to be related to infliximab. For placebo, SAEs occurred in 130 participants (25.2%) and the events were attributed to trial product in 7 of these participants (1.4%). One participant in the infliximab group experienced a grade 1 (mild) infusion reaction. Safety assessment is presented in Table 3.

The percentage of participants who had any secondary infection was similar at day 60 with infliximab and placebo (79 [15.3%] vs 72 [14.0%]) (Table 3). The most common secondary infections were bacterial pneumonia, bloodstream, and urinary tract infections (eTable 18 in Supplement 2).

The ACTIV-1 IM master protocol evaluated the benefit of an additional immunomodulator added to standard care in hospitalized participants with moderate to severe COVID-19 pneumonia. Within this platform, 3 agents with distinct mechanisms of action—abatacept, cenicriviroc, and infliximab—were evaluated. The ACTIV-1 study found that none of the substudies reached a statistically significant primary end point of time to recovery. Thus, 28-day mortality and day 14 clinical status cannot be considered statistically significant based on the predefined gatekeeping approach, despite nominally significant CIs, although these end points are clinically important. For abatacept and infliximab substudies, the primary and 2 key secondary end points were all in the beneficial direction, with the mortality benefit showing remarkable consistency across subgroups (eFigures 3 and 13 in Supplement 2). Additionally, the consistent lack of benefit among all 3 outcomes for cenicriviroc within the same master protocol further supports the validity of the findings for abatacept and infliximab because it illustrates differences in outcomes between the 3 drugs despite each drug being compared with placebo groups that were similar (although not identical).

There is a growing body of evidence for use of a second immunomodulatory agent in addition to corticosteroids. Positive results from several trials prompted the NIH treatment guidelines to include baricitinib and tocilizumab^5,6,9,18 as a second immunomodulator in addition to corticosteroids for patients with progressive respiratory failure and evidence of systemic inflammation.¹⁹ The current results offer additional information on the efficacy and safety of abatacept and infliximab. Although severe disease is less common in the postvaccination era and SARS-CoV-2 variants become less deadly, immunomodulatory therapy is still critical. The US Centers for Disease Control and Prevention continued to report more than 1000 COVID-19–associated deaths per week in the US as of April 2023.²⁰ Data on the benefit of additional immunomodulators is particularly important in the setting of previous drug shortages within these categories during periods of the pandemic. Additionally, options that are safer for patients with kidney, liver, or clotting disorders fill important gaps.

Similar mortality benefits have been demonstrated for a number of immunomodulatory agents with varying mechanisms (IL-6 inhibition, Janus kinase-1/2 inhibition, TNF inhibition, and select T-cell costimulatory modulation), although further exploration of the role of inflammation in the pathogenesis of SARS-CoV-2 is needed because modulation of other pathways, such as those impacted by cenicriviroc, does not show benefit. The term cytokine storm describes the marked increase in cytokines that triggers a cascade of synergistic inflammatory signaling in severe disease.² Aberrant T-cell function in severe COVID-19 is complex, involving hyperactivation of CD4+ T cells.¹ Circulating monocytes and macrophages also play a key role in inflammatory signaling and fibrosis.²¹ Subtleties in ACTIV-1 substudies hint at small differences in therapeutic roles. For example, subgroup analysis suggested a larger effect for abatacept among older adults (≥65 years old) and those with diabetes than observed for other subgroups in the abatacept substudy (eFigures 1 and 3 in Supplement 2) or the effects found for the infliximab substudy (eFigures 12 and 13 in Supplement 2). Deeper exploration of dissimilarities among mechanisms and effects of these drugs on the COVID-19 inflammatory cascade could improve insight into the pathogenesis of severe disease.

A key clinical question involves how to treat hospitalized patients receiving low-flow oxygen. Current NIH guidelines do not recommend addition of a second immunomodulator for these patients in the absence of rapidly increasing oxygen needs. The ACTIV-1 IM substudies begin to address this critical question, showing reduced mortality for participants receiving abatacept or infliximab in a general population of participants of all ordinal scales independent of inflammatory markers. Subgroup analysis suggests that the benefits in mortality occur for participants receiving both low-flow and high-flow oxygen/noninvasive ventilation (eFigures 3 and 13 in Supplement 2).

The ACTIV-1 IM master protocol provides a helpful blueprint for a high-quality trial design that allows for concurrent evaluation of multiple agents while simultaneously providing a shared placebo group. Other COVID-19 platform trials conducted provide shared controls, but often do not offer concurrently enrolled controls or a placebo group.^4-6,22 Use of a shared placebo represents a patient-centered approach to reduce the number of patients receiving placebo, allowing adequate power to evaluate multiple agents with a smaller number of participants enrolled.¹¹ The findings that active and shared placebo groups had similar baseline characteristics, and that there was no benefit seen when cenicriviroc was compared with the shared placebo while benefits were observed for abatacept and infliximab, reinforce the integrity of the shared placebo group in this study.

This study highlights a major challenge in the selection of a primary end point for this and other COVID-19 studies in the setting of a rapidly changing clinical arena.²³ A mortality primary end point is considered definitive and free from subjectivity, but can require large participant enrollment potentially resulting in prolonged study recruitment with delayed results. In addition, mortality does not encompass other patient-centered outcomes, which are a greater focus in end points such as time to recovery or clinical status. As research moves forward for COVID-19, the selection of primary end points will be important to encompass evaluable patient-centered outcomes, but take steps to do so with lower enrollment targets.

As SARS-CoV-2 moves from being pandemic to endemic, and while new variants continue to emerge, ongoing morbidity and mortality from this disease is likely. Variants may evade specific antiviral strategies, but treatments that augment host responses are likely to be more durable. Expanding treatment armament available globally and developing optimized treatment strategies remains paramount.

The ACTIV-1 study has several limitations. First, ACTIV-1 took place prior to the Omicron era, raising the question of whether the results are limited to the variants represented at that time. However, despite changes in the predominant circulating variant over time and decreased disease severity overall, patients continue to be admitted to the hospital with respiratory failure and evidence of immune dysregulation, highlighting the ongoing need for optimized treatment strategies. Second, although vaccinations became available during the course of the study, data were not available on participant vaccination status. Third, unlike abatacept and infliximab, which were conveniently dosed as 1-time intravenous infusion, cenicriviroc is an oral medication that was administered in this study twice daily for 28 days. This led to challenges with adherence, which may have contributed to the negative results observed in this substudy, although similar adherence issues would also exist in a clinical setting.

Time to recovery from COVID-19 pneumonia among hospitalized participants was not significantly different for abatacept, cenicriviroc, or infliximab vs placebo.

Accepted for Publication: June 6, 2023.

Published Online: July 10, 2023. doi:10.1001/jama.2023.11043

Corresponding Author: William G. Powderly, MD, Division of Infectious Diseases, Department of Medicine, Washington University School of Medicine, 4990 Children's Pl, St Louis, MO 63110 (wpowderly@wustl.edu).

Author Contributions: Dr Powderly had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs O’Halloran and Ko contributed equally to this report.

Concept and design: O’Halloran, Ko, Anstrom, McCarthy, Panettieri, Lachiewicz, Gonzalez, Patelli Lima, Blamoun, Alicic, Wolfe, Maldonado, Ferguson, Halabi, Benjamin, Adam, Chang, LaVange, Proschan, Bozzette, Powderly.

Acquisition, analysis, or interpretation of data: O’Halloran, Ko, Anstrom, Kedar, McCarthy, Panettieri, Maillo, Segura Nunez, Lachiewicz, Gonzalez, Smith, Mendivil-Tuchia de Tai, Khan, Mena Lora, Salathe, Capo, Rodríguez Gonzalez, Patterson, Palma, Ariza, Blamoun, Nannini, Sprinz, Mykietiuk, Alicic, Rauseo, Wolfe, Witting, Wang, Parra-Rodriguez, Der, Willsey, Wen, Silverstein, O’Brien, Al-Khalidi, Maldonado, Melsheimer, Ferguson, McNulty, Zakroysky, Halabi, Benjamin, Butler, Atkinson, Chang, LaVange, Proschan, Bozzette, Powderly.

Drafting of the manuscript: O’Halloran, Ko, Kedar, McCarthy, Panettieri, Lachiewicz, Khan, Mena Lora, Patelli Lima, Blamoun, Wolfe, Wang, Melsheimer, McNulty, Halabi, LaVange, Bozzette, Powderly.

Critical revision of the manuscript for important intellectual content: O’Halloran, Ko, Anstrom, Kedar, Panettieri, Maillo, Segura Nunez, Lachiewicz, Gonzalez, Smith, Mendivil-Tuchia de Tai, Khan, Mena Lora, Salathe, Capo, Rodríguez Gonzalez, Patterson, Palma, Ariza, Blamoun, Nannini, Sprinz, Mykietiuk, Alicic, Rauseo, Witting, Parra-Rodriguez, Der, Willsey, Wen, Silverstein, O’Brien, Al-Khalidi, Maldonado, Ferguson, Zakroysky, Benjamin, Butler, Atkinson, Adam, Chang, LaVange, Proschan, Bozzette, Powderly.

Statistical analysis: Anstrom, Kedar, McCarthy, Panettieri, Wolfe, Wen, Silverstein, O’Brien, Al-Khalidi, McNulty, Zakroysky, Halabi, LaVange, Proschan, Bozzette.

Obtained funding: Patterson, Benjamin, Adam.

Administrative, technical, or material support: Maillo, Gonzalez, Smith, Mendivil-Tuchia de Tai, Mena Lora, Salathe, Capo, Patterson, Palma, Blamoun, Mykietiuk, Rauseo, Witting, Wang, Parra-Rodriguez, Maldonado, Melsheimer, Ferguson, Butler, Atkinson, Adam, Chang, Bozzette.

Supervision: Gonzalez, Mendivil-Tuchia de Tai, Salathe, Capo, Rodríguez Gonzalez, Palma, Blamoun, Benjamin, Butler, Chang, Bozzette, Powderly.

Other - Recolección de participantes, evaluación, monitoreo: Segura Nunez.

Conflict of Interest Disclosures: Dr O’Halloran reported receiving grants from Janssen Scientific outside the submitted work. Dr Anstrom reported receiving grants from Merck, Bayer, NIH, and Pfizer during the conduct of the study. Dr Panettieri reported receiving grants from AstraZeneca, RIFM, TEVA, Medimmume, AgoMab, ACTIV-1, Janssen, and Vault Health; personal fees from Sanofi, Merck & Co, and AstraZeneca; and serving on an advisory beard for Genentech and RIFM outside the submitted work. Dr Maillo reported receiving institutional investigation fees from Technical Resources International, Inc during the conduct of the study. Dr Lachiewicz reported receiving funding for clinical trial participation from Duke Clinical Research Institute during the conduct of the study; clinical trial research support from Pfizer, Shionogi, Novartis, Cidara, and Contrafect outside the submitted work. Dr Smith reported receiving grants from BARDA during the conduct of the study and personal fees from UCB and Syneos Health outside the submitted work. Dr Mendivil-Tuchia de Tai reported receiving sponsorship for Technical Resource International during the conduct of the study. Dr Khan reported receiving grants from Roche, Dompe Pharmaceuticals, United Therapeutics, Baxter, Regeneron, and Johnson & Johnson and personal fees from Dompe Pharmaceutics outside the submitted work. Dr Mena Lora reported receiving grants from NIH during the conduct of the study. Dr Salathe reported receiving grants from NIH during the conduct of the study and personal fees from Arrowhead Pharmaceuticals; collaborations with Enterprise Therapeutics; and grants from NIH, FAMRI, and Cystic Fibrosis Foundation outside the submitted work. Dr Rodríguez Gonzalez reported receiving grants from Technical Resources International TRI during the conduct of the study and grants from Regeneron Pharmaceutical, Inc outside the submitted work. Dr Patterson reported receiving grants from Gilead to UT Health San Antonio outside the submitted work. Dr Ariza reported having a contract with Technical Resources International funded by BARDA during the conduct of the study. Dr Nannini reported receiving grants from Pfizer and Basilea and personal fees from GSK and MSD outside the submitted work. Dr Wolfe reported receiving personal fees from Gilead, Regeneron, and Adagio and serving on a board for Biogen, Adamis, and Atea outside the submitted work. Dr O’Brien reported receiving grants from Technical Resources International/Biomedical Advanced Research and Development Authority during the conduct of the study. Dr Maldonado reported receiving personal fees from Bristol Myers Squibb during the conduct of the study. Dr Melsheimer reported being an employee of and stockholder in Janssen Pharmaceuticals, a division of Johnson & Johnson, which commercializes one of the drugs, infliximab, evaluated in the ACTIV-1 study. Dr Ferguson reported being an employee of AbbVie during the conduct of the study. Dr Halabi reported being a member of a data and safety monitoring board for Sanofi, Aveo, Janssen, and Bristol Myers Squibb during the conduct of the study. Dr Benjamin reported providing consultancy for AbbVie consultancy, PPD, and Syneos Health outside the submitted work. Dr Butler reported contract from BARDA (NCATS Technical Lead) during the conduct of the study and contracts with NIH outside the submitted work. Dr Adam reported receiving US Government funding through OWS during the conduct of the study. Dr Bozzette reported receiving salary from NIH during the conduct of the study. Dr Powderly reported receiving grants from National Center for Advancing Translational Sciences and Technical Resources International Support for Role as ACTIV-1 PI during the conduct of the study and personal fees from Merck Laboratories outside the submitted work. No other disclosures were reported.

Funding/Support: The study was funded with federal funds from the US Department of Health and Human Services, Office of the Assistant Secretary for Preparedness and Response, Biomedical Advanced Research and Development Authority, under contract number HHSO100201400002I/75A50120F33002. Infrastructure support and resources for research reported in this publication were provided in part by NCATS of the National Institutes of Health under award number(s): UL1TR002345 (Washington University in St Louis); U24TR001608 (Duke University–Vanderbilt University Medical Center Trial Innovation Center [Trial Innovation Network]); U24TR001579 (Vanderbilt University Medical Center Recruitment Innovation Center); UL1TR001453 (University of Massachusetts Medical Center); UL1TR003017 (Rutgers New Jersey Medical School); UL1TR003107 (University of Arkansas Medical Sciences); UL1TR002645 (University of Texas Health Science Center at San Antonio); UL1TR001430 (Boston Medical Center); UL1TR001881 (UCLA Medical Center); UL1TR002366 (University of Kansas Medical Center); UL1TR003096 (Tulane University Health Science Center); UL1TR002489 (UNC-Chapel Hill); UL1TR001998 (University of Kentucky); UL1TR002553 (Duke University Medical Center); UL1TR002377 (Mayo Clinic); UL1TR002544 (Tufts Medical Center); UL1TR002319 (University of Washington Medical Center); UL1TR001445 (NYU Langone Health/Tisch Hospital).

Role of the Funder/Sponsor: Janssen provided infliximab, Bristol Myers Squibb provided abatacept, and AbbVie provided cenicriviroc for use in this trial but did not provide any financial support. Gilead Sciences provided remdesivir for use in this trial but did not provide any financial support. Employees of Janssen and Gilead Sciences participated in discussions about protocol development and in weekly protocol team calls. The final trial protocol was developed by the protocol chair, Dr William Powderly, the IND sponsor Dr Daniel K. Benjamin Jr, and a protocol development committee including representatives from the National Center for Advancing Translational Sciences (NCATS). NCATS had a collaborative role in the trial design, management, interpretation of the data and the preparation of the manuscript along with the protocol chair, the study statisticians and the study writing group.

Group Information: The ACTIV-1 IM Study Group Members are provided in Supplement 3.

Disclaimer: The findings and conclusions in this publication are those of the authors and do not necessarily represent the views of the Department of Health and Human Services or its components.

Data Sharing Statement: See Supplement 4.

Additional Contributions: We thank the members of the ACTIV-1 Study Team (see Supplement 3) for their many contributions in conducting the trial, the members of the data and safety monitoring board, William Blackwelder, PhD (chair) (University of Maryland School of Medicine, Baltimore, MD); Timothy G. Buchman, PhD, MD (Emory University School of Medicine, Atlanta, GA); Wilbur H. Chen, MD, MS (University of Maryland School of Medicine, Center for Vaccine Development and Global Health, Baltimore, MD); Lawrence H. Moulton, PhD (Johns Hopkins University, Bloomberg School of Public Health, Baltimore, MD); David M. Parenti, MD (The George Washington University School of Medicine, Washington, DC); Carol O. Tacket, MD (University of Maryland School of Medicine, Center for Vaccine Development and Global Health, Baltimore, MD); William Checkley, MD, PhD (Johns Hopkins University, Baltimore, MD); and Beatriz Grinsztejn, MD (Instituto Nacional de Infectologia Evandro Chagas-Fiocruz, Rio de Janeiro, Brazil), for their oversight; the community advisory board, Larisa Caicedo, Ashish Cowlagi, Anna Davis, Lincoln Larmond, Doug Lindsey, Bob Pearson, and the participants and their families for their altruism in participating in this trial; Elizabeth E.S. Cook for editorial support; and Christopher J. Lindsell for contributions to the statistical analysis.