They teased us with a press release weeks ago, and Faucci has been promoting it ever since – and now, we finally have the data from the ACTT-1 trial on remdesivir. Here we take a closer look to see if this really is the "very promising result" it was sold as.
Reminder: background on remdesivir
- Antiviral with in-vitro promise
- Recent trial from Wang et al1 was negative for a difference in time to clinical improvement
- Trial was underpowered as it was stopped early due to slow recruitment
- They did note a trend towards faster improvement (18d vs 23d) in the subset of patients who received remdesivir within 10d of symptom onset
ACTT-1: Adaptive COVID-19 Treatment Trial2
- Randomized, double-blind, placebo-controlled trial of remdesivir for COVID-19
- Enrollment for 3 weeks in Feb/March at 60 sites in the US, Europe, and Asia
- Ended up being predominantly North American patients (about 80%)
- Methods
- Patients: adults with PCR-confirmed COVID infection with at least one of the following: radiographic infiltrates, SpO2 < 94% on RA, need for supplemental oxygen, or MV/ECMO
- Sub-stratified further into mild/moderate (SpO2 > 94% on RA, RR <24) vs severe disease; most patients (about 88%) had "severe" disease
- Key exclusion: transaminitis (AST/ALT > 5x ULN), renal dysfunction (GFR < 30), pregnancy/breastfeeding
- Intervention: remdesivir 200mg IV on day 1, then 100mg IV daily on days 2-10
- Primary outcome: time to recovery, defined as the first day that the patient scored 1, 2, or 3 on an 8-point clinical scale
- 1 = totally well; 3 = still hospitalized but not requiring oxygen or ongoing medical treatments
- Primary outcome was actually changed near the start of the trial (previously designed as change in 8-point ordinal scale on day 15)
- Changed due to evolving knowledge of the potentially protracted course of COVID-19
- Secondary outcomes: 14d and 28d mortality; as well as adverse events
- These results are actually an interim analysis: 1/3 of patients are still enrolled in the trial and have no documented final outcome yet
- Because of the nature of the results, the data safety monitoring committee recommended the results be released while the trial is still ongoing
- Results
- Patient population
- 1063 patients enrolled; though 301 are still continuing in the trial follow-up period (but their data to date was included in the analysis)
- Relatively sick; most people were hospitalized requiring at least supplemental oxygen; about 25% of patient were mechanically ventilated or on ECMO
- Most had at least one co-morbidity; most commonly hypertension (present in almost 50% of patients)
- Primary outcome
- Remdesivir group has faster time to clinical recovery: 11d vs 15d (rate ratio for recovery 1.32, CI 1.12-1.55)
- Note that the sickest subgroups – those on non-invasive or invasive ventilation, or ECMO – did not demonstrate any benefit
- Perhaps because in these groups there is a significant immune hyperactivation component?
- No change in the significance of the outcome when they adjusted for those randomized <10d after symptom onset vs >10d
- Secondary outcomes
- Mortality at 14d was numerically lower in the remdesivir group, but not (quite) statistically significant – HR for death 0.70 (CI 0.47-1.04)
- No significant difference in adverse events
- Caveats
- Change in primary outcome: we never like to see this in trials, but seems to have been done here for a fairly legitimate indication as our understanding of this disease evolved
- Notably, the analysis for their original primary endpoint still shows a statistically significant benefit to remdesivir
- Early unblinding and reporting of the trial: done when a planned interim analysis revealed a clinically significant benefit to remdesivir. The committee made the reasonable decision that in a pandemic situation, these were important results to be released as soon as possible.
- The authors themselves note that it is important that the full trial period (with follow-up and monitoring) be completed, and these full results will be reported separately
- It's interesting that they chose to report mortality at 14 days, when their change in primary outcome acknowledges that COVID often has a more protracted course and requires a longer time frame of study. Although the reported difference sounds numerically impressive, the Kaplan-Meyer curves appear to be converging, suggesting that the benefit is only transient
- A lot of granular data is missing from this manuscript – hopefully will be included in the follow-up paper
Compared to the Wang trial
- Overall these trials are fairly congruent (though one was technically negative, and one was positive)
- Both show a reduced in time to clinical improvement with remdesivir
- Statistical significance of the ACTT-1 trial (in contrast to Wang) likely primarily due to higher numbers, more statistical power, and perhaps the use of an 8-point vs 6-point ordinal scale
- Not convincing effect on mortality in either trial
- The ACTT-1 group reports that they measured viral load (as did Wang, who found no difference), but it's not included in this paper
Where this leaves us
- This is potentially promising!
- Remdesivir appears to hasten time to recovery - even without any effect on mortality, this could have very significant implications from a resource utilization perspective, especially in a pandemic
- BUT note that it is an IV drug - do we really need everyone hospitalized for 10 days to get the full treatment protocol?
- Mortality benefit is unclear; disappointingly, the curves seem to come together
- Remdesivir does appear safe
- We need the full data – stay tuned!
Sources
- Wang Y, Zhang D, Du G et al. Remdesivir in adults with severe COVID-19: a randomised, double-blind, placebo-controlled, multicentre trial. Lancet, 2020;395: 1569-78. doi:10.1016/S0140-6736(20)31022-9
- Beigel JH, Tomashek KM, Dodd LE et al. Remdesivir for the Treatment of Covid-19 — Preliminary Report. New Eng J Med, May 2020. doi:10.1056/NEJMoa2007764
