Thread by @FranciscoMarty_

Updated the #RECOVERYtrial topline results summary table of the #dexamethasone arm compared to usual care for treatment of hospitalized patients with #COVID19.

Got additional information colleagues shared in Twitter.

Thread with some comparisons with #ACTT1

#IDtwitter 1/

the #ACTT1 authors have only reported (preliminary) Day 14 mortality, and the #RECOVERYtrial has only posted Day 28 mortality.

But these days the Supplementary Appendices in trial reports have additional relevant information:
nejm.org/doi/full/10.10…

#DontCoverDiscover

The paper says they were not reporting day 29 mortality, but If you look at Figure S3 in the #NEJM #ACTT1 preliminary report, you have KM estimates out to day 30.

From there you can visualize the probability of death at day

28 (see figure) have left lines and scales

You can see above that overall estimated Day 28 mortality in #ACTT1 was ~11.8% for patients randomized to #remdesivir compared to ~14.9% for those who received #placebo.
Caveat is that ~30% of the study patients had not completed follow up to day 28, curves flatten after day 21

But in the #NEJM #ACTT1 supplemental Figure S3, you are also given the KM estimates of survival/mortality for the #remdesivir and placebo arms based on respitarory needs at randomization. See 3 figures here.
Population on high-flow/non-invasive ventilation small, so not included

So with this estimates from #ACTT1 #remdesivir trial you can make some more quantitative comparisons with he #RECOVERYtrial #dexamethasone arm for day 28 mortality. as I tabulated here.

Interesting stuff to ponder. Some space for you to think about it...

Obvious and striking is the difference in 28 day mortality between the control arms, with the mortality being twice higher in the #RECOVERYtrial compared to #ACTT1.

#ACTT1 did restrict patients with significant renal dysfunction and high LFTs but may not explain it all.

The #RECOVERYtrial claims a highly significant test for trend between the 3 groups, (see summary table) and in the figure kindly posted by .@drnickeasom.

But not sure if a trend test is a appropriate when you go from protection to harm (cross RR=1)

https://twitter.com/drnickeasom/status/1272913866389893121?s=20

very interesting to see the large absolute & relative benefit among #COVID19 patients who required supplemental O2 who received #remdesivir compared to similar group who received #dexamethasone
May be better to use #remdesivir first in this severity band first if you have access

For patients with #COVID19 who need mechanical ventilation, I think the #RECOVERYtrial is quite compelling on the benefit of low-dose #dexamethasone.
The #ACTT1 KM curve shows a small #remdesivir benefit initially, but then the curves cross.

Can't wait for the final report!

Some may disagree that Day 28 comparisons are fair as the #ACTT1 data has not been finalized. I don't think the final estimates will be that different from those in the supplement.
I also did a comparison of Day 14 KM estimates between trials in this table, which are complete

The #ACTT1 numbers come from Table 2 of the #NEJM preliminary report.

The Day 14 KM mortality estimates for the #RECOVERYtrial were done using graphs from a source that remains anonymous, but all data numbers correspond to what has been posted online.