Thread by @FranciscoMarty_

Posting some reflections on the #ACTT & #RECOVERY trials related to #remdesivir & #dexamethasone for the treatment of #COVID19

Those results feel like definitive advances, but we need to realize their limitations and need for confirmation despite everyone's #pandemic fatigue

1/

There are lot of skeptics in and outside of #Twitter on the results of both trials, strong emotions regarding big-pharma made medications, the way the #EUA was made, also about use of #steroids in the #ICU that has a quite rocky history among those who run those units.

And we need to acknowledge from the outset all sorts of basic, and not so basic, emotions that come in the doing clinical trials, exacerbated by academic cliques, national defense issues (and nationalisms) that can get exacerbated with the strains in health systems.

First hit the (preliminary) results of the #ACTT1 trial, meeting their primary endpoint of a faster time to #Recovery...
Wasn't RECOVERY the name of the other trial, how come it is the outcome of #ACTT?...

We had rarely used ordinal scales before in Infectious Diseases...

It fell like like a #shellshock to many as the #EUA in the US was approved and #remdesivir became available even before we were able to see final results and what does recovery mean anyway?
The hardliners wanted less deaths for sure, but the study was not designed that way.

The #RECOVERY came to the rescue. Large simple trial, lots of patients. #Hydroxychloroquine does not work, but #dexamethasone did! A big win for low-cost, widely available treatments fans.

But both the #ACTT1 and #RECOVERY trial results have perplexing nuances:
- In #RECOVERY, #dexamethasone helped nicely those on #ventilators, but #remdesivir had no effect.
- those on supplemental oxygen, #remdesiver worked beautifully, but the #dexamethasone effect was small

Discouraging was that there was a signal for harm among patients who were hospitalized and did not require oxygen who got #dexamethasone, a big flag in my mind despite any handwaving you want to do, and the effect of #remdesivir was notable, but small given the low placebo events

There were no major effect modifications in the #remdesivir data other than baseline status at randomization, but #dexamethasone was interesting: those younger than 70 who were sick for over 7 days seemed to benefit the most, men more than women. #RECOVERY

#RECOVERY was open label and most importantly, it was not stratified by center, knowing from prior experience and now with data, that care is strongly influenced by local practices and systems being overwhelmed.

BUT sample size was huge for ID trials, so the answer was "#Robust"

So let's take a step back and think the pros and cons of both trials, what we can learn from them and how can we inform and probably confirm these results.

First a look at the #design of the #clinicaltrials (trying some #emoji scales here)

Besides the obvious issues of #bias with open label, non-placebo controlled issues, the main concern from a technical perspective is that both #RECOVERY and #ACTT1 claim effects on different strata, but those strata were selected post-hoc, were not stratified at #randomization

At such, wether #robust or not, these considerations need to be considered #hypothesis generating and should require confirmation in a new trial.

As you make additional comparisons, some other issues become apparent.
#RECOVERY had no limitations in organ disfunction, but I don't think we can attribute the 2xmortality rates solely to that.
#RECOVERY didn't specify an imaging diagnosis of lung disease or confirmation of SARS

The #mortality data for #ACTT1 is posted in their appendix.
Given the number of patients who had finished at the time of their preliminary report, I don't expect sizable changes in their final report, so I decided to compare the mortalities head-to-head with #RECOVERY.

Used the #RECOVERY mortality Y-axis for the comparisons and adjusted the #ACTT1 graphs to scale. Also putting the key numbers in.
The absolute overall death difference is ~3% in both trials, not too exciting if reading from afar.
For patients on #ventilators, dex worked nicely

In patients who required supplemental #oxygen, #remdesivir worked amazingly well, dexamethasone was robustly weak in its effect.

Many doctors are already giving both to patients, but are we helping them by doing this? Is the combination synergistic, indifferent, antagonistic?

In those who did not require oxygen at baseline, which we would call hospitalized with #moderate #COVID19, dexamethasone not helpful and remdesivir seemed to have an small impact on mortality.

So you can see how different people can be swayed to use #remdesivir +/- #dexamethasone in different situations:

- would remdesivir benefit patients on the ventilator further if used with dexamethasone?
- would dexamethasone antagonize the benefit of remdesivir in those on O2?

I think we seriously need to consider a new clinical trial to confirm the subgroup analyses of #ACTT1 & #recovery

- For the optimists, may call it the #CoVfirm trial
- For the pessimists, may call it the #CoVince-ME trial

But here is the proposal for the #ReDexCoVer trial

Maybe .@DrTedros .@WHO .@JeremyFarrar .@wellcometrust .@BillGates .@gatesfoundation, other parties can can think this through and get it done.