RECOVER have put out how they chose the biomarkers for their Long COVID trials and it's some of the most uninspired and methodologically questionable work I've seen in this space.
They've got over a billion in funding. And what they did was go through other people's studies, look for blood markers that came up abnormal in at least two of them, and use that as their list. So we've got 82 markers, and honestly the reason most of them are in there is because they're cheap to measure, not because they're the right markers.
The studies they took them from are mostly small and have most of the limitations Long COVID studies always have. Heterogeneous patient populations, no subgrouping, poor inclusion criteria, no mapping for severity, etc etc. Plus huge batch effects across all the different studies. You can't just pool that together and call it evidence.
It gets worse. They only prioritised studies that compared patients to people who'd never had COVID. Researchers from major studies have already pointed out that when you do that, you can't separate Long COVID from leftover inflammation from the infection. So their filter pushed out the better designed studies.
And complement, one of the main signals coming out of the larger studies, isn't on their list at all. Not one complement protein in 82 markers.
Then there's nothing on their page about how they'll deal with batch effects in their own trials either. This stuff should be centralised, not left to individual sites to run assays.
They also removed any markers that weren't "feasible" to measure at trial sites. That's just not true. You can do almost anything nowadays with specialist tubes. RNA tubes, DNA tubes. Process serum and plasma on site and ship it. Isolate PBMCs on site, freeze them and send them on dry ice. All of this exists already. They could have done it for everything they dropped. They had a billion.
Long COVID isn't one thing, we all know there are subgroups. Instead they're comparing people with Long COVID to healthy controls on average, which just washes out the differences between us.
RECOVER are making ad hoc decisions without thinking them through critically, because they have no skin in the game. It's poor, half baked decisions with zero thought in them, again and again, just to look like they're doing something. They're chasing their tail, and in doing so they're wasting all of our money. They desperately need to take a step back and create a cohesive long term strategy.
This is just going to lead to another year, and another press conference saying sorry, our work last year was useless and it failed.
RECOVER-TLC has published a list of 82 blood biomarkers to be measured in its Long COVID treatment trials.
It's not a diagnostic test, and it won't change anything in the clinic for now. But it's worth looking at what's on the list, and especially what isn't🧵