VIP (Vasoactive Intestinal Peptide)

The average time to discharge was 7.8 days on aviptadil and 10 days on placebo (P=0.049). Breathlessness on the modified Borg scale was not significantly different on day 3 (P=0.090) but was lower on aviptadil on day 7 (P=0.033). The chest CT lung-damage score improved more on aviptadil by day 28 (P=0.028). By day 180, 2 patients had died on aviptadil and 5 on placebo (5.1% versus 12.2%), a difference that was not statistically significant (P=0.433).

• A small Phase 2 trial (80 patients); the authors say that, because of drop-outs, the number who completed the study was too low to generalize.
• The paper's primary outcome (time to discharge) differs from the registry's primary outcome (clinical improvement), and the sources do not explain why.
• The difference in deaths (2 versus 5 patients) was not statistically significant.
• Some oxygen-saturation readings were taken while patients were on supplementary oxygen, which affects the averages.
• It enrolled hospitalized patients who did not need intensive care at admission, so results do not apply to critically ill patients.
• The study was funded by the drug's developer, Centurion Pharma, and the registry posts no results, so numbers could not be cross-checked.

• The trial's primary outcome: ClinicalTrials.gov reports Primary: clinical improvement (6 months); time to discharge listed as a secondary outcome; Medical Principles and Practice paper reports Primary: time from hospitalization to discharge within 30 days; clinical and radiological scores secondary. The paper's main outcome is a registry secondary outcome. The reason for the difference is not established from the sources reviewed.

Aviptadil versus placebo: the odds of a better day-90 status were not significantly different (odds ratio 1.11, 95% CI 0.80 to 1.55, P=0.54). By day 90, 86 of 231 people on aviptadil (38%) and 83 of 230 on placebo (36%) had died (hazard ratio 1.04, 95% CI 0.77 to 1.41, P=0.78). An early safety measure covered serious or grade 3-4 events, organ failure, serious infection or death through day 5. It occurred in 146 of 231 (63%) on aviptadil and 129 of 230 (56%) on placebo (odds ratio 1.40, 95% CI 0.94 to 2.08, P=0.10). The aviptadil comparison was stopped for futility. Remdesivir versus placebo: odds ratio 0.98 (95% CI 0.45 to 2.11, P=0.96); 17 of 44 people on remdesivir and 20 of 43 on placebo died by day 90 (hazard ratio 0.74, 95% CI 0.39 to 1.42, P=0.37). That comparison ended early because of slow enrollment, and the journal abstract says it did not permit definitive conclusions.

• The aviptadil comparison was stopped early for futility and the remdesivir comparison was stopped because of slow enrollment, leaving it far smaller than planned (87 people); the journal abstract says it did not permit definitive conclusions.
• The two drugs were tested in a factorial and strata design in which only 85 people were randomized to both; most aviptadil participants were in a stratum where remdesivir was already started or not an option, so the results cannot isolate the combination.
• Participants were critically ill (94% in intensive care) with COVID-19 respiratory failure during 2021-2022; results may not apply to milder illness or to later virus variants and treatments.
• Of the 471 people randomized to aviptadil or placebo, 461 who received at least part of an infusion were analyzed, and 11 had no day-90 status for the primary outcome.
• The main ClinicalTrials.gov record lists only cohort counts; this summary combines it with two substudy records and the journal abstract.
• The full journal paper was not reviewed in full.

The main goal was not met. At day 60, 72 of 131 people (55.0%) on aviptadil and 34 of 65 (52.3%) on placebo were alive and free of respiratory failure; the prespecified analysis, which accounts for the baseline ventilation imbalance, gave an odds ratio of 1.6 (95% confidence interval 0.86 to 3.11), not statistically significant. Day-60 survival, a key secondary outcome, was higher with aviptadil: 46 of 131 (35.1%) had died versus 30 of 65 (46.2%) on placebo (odds ratio 2.0, 95% confidence interval 1.05 to 3.87, P=0.035). A subgroup finding in people who started on a ventilator was based on small numbers and the paper calls such results hypothesis-generating.

• The final primary outcome was not statistically significant.
• The primary and key secondary outcomes were swapped partway through the trial, before unblinding, which complicates how to read the survival finding.
• The groups were not evenly matched at the start: 55 of 131 people on aviptadil and 14 of 65 on placebo were on mechanical ventilation.
• Subgroup results, including the 10-fold survival odds in ventilated patients, come from small groups and are hypothesis-generating.
• The registry and the paper give slightly different counts for the main outcome and label different time points, and the paper's text does not match its own table.
• The standard of care was not set by the protocol and could differ between hospitals, and the paper notes that COVID-19 surges strained community hospitals during the study.
• Participants all had critical illness with respiratory failure, so results do not apply to milder COVID-19.

• People alive and free of respiratory failure (main outcome): ClinicalTrials.gov posted results (labelled day 28) reports 72 of 131 on aviptadil; 33 of 65 on placebo; Critical Care Medicine paper, day-60 outcomes table reports 72 of 131 (55.0%) on aviptadil; 34 of 65 (52.3%) on placebo. The aviptadil counts match, but the placebo count differs by one person and the registry labels the time point day 28 while the paper's day-28 table shows different counts. The reason for the difference is not established from the sources reviewed.
• Which outcome was the trial's primary outcome: Original protocol (as described in the paper) reports Survival was primary and recovery from respiratory failure was key secondary; Final analysis plan (as described in the paper) reports Alive and free of respiratory failure at day 60 was primary; day-60 survival was key secondary. The two outcomes were swapped in February 2021, at the data monitoring committee's suggestion and before the blinded data were unblinded, according to the paper. The paper states the swap happened before unblinding and that it was done at the data monitoring committee's suggestion. This summary reports the final primary outcome and also shows survival.