Review Negative results in long COVID clinical trials: choosing outcome measures for a heterogeneous disease, 2025, de Canson et al

[John Mac]

Long COVID is an umbrella term for the heterogeneous long-term consequences of SARS-CoV-2 infection. It encompasses a wide spectrum of phenotypes (sometimes overlapping), and its underlying mechanisms might vary across and within phenotypes. Trials that do not account for this heterogeneity risk false-negative results. First, endotypic heterogeneity might produce heterogeneity in treatment response, and inadequate methodologies risk eclipsing responder subgroups. Second, and the focus herein, phenotypic heterogeneity constitutes a challenge for the choice of outcome measures. Patient-reported outcome measures (PROMs) are essential elements in trial design (and key to surrogate biomarker validation), as they measure clinically relevant outcomes directly. However, long COVID's considerable phenotypic heterogeneity makes it difficult to identify PROMs that can sensitively assess dimensions of disease severity across manifestations while maintaining statistical power. Several highly anticipated long COVID clinical trials, such as PAX-LC,1 used PROMs that might not have full validity in long COVID, potentially contributing to the negative results of these trials

https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(25)00665-6/abstract?rss=yes

 

[Utsikt]

Trials that do not account for this heterogeneity risk false-negative results. First, endotypic heterogeneity might produce heterogeneity in treatment response, and inadequate methodologies risk eclipsing responder subgroups. Second, and the focus herein, phenotypic heterogeneity constitutes a challenge for the choice of outcome measures. Patient-reported outcome measures (PROMs) are essential elements in trial design (and key to surrogate biomarker validation), as they measure clinically relevant outcomes directly.

So they’ve completely missed the point even though it’s right in front of them in their own text. Inclusion criteria are far more important than PROMs.

And they seem very keen on using PROMs without discussing the important limitation that is frequent lack of blinding.

They also ignore objective proxies like step count, hours worked/studied, etc. that could do just as well as PROMs, if not even better.

It’s nice that FUNCAP gets some recognition for ME/CFS, though, but I don’t understand how you’d go about testing if FUNCAP works as an outcome in a trial when we don’t have any effective treatments to test it with. The only way to validate that FUNCAP works would be to see if it matches up with the changes in either biomarkers or objective outcomes.

I guess you could test if it produces false positives, but that’s the opposite of avoiding false negatives that they are concerned about.

 

[Jaybee00]

Some interesting affiliations

Department of Theoretical Philosophy, University of Groningen, Groningen 9712 GL, Netherlands

e New York, NY, USA

Are there non-theoretical philosophy departments?

 

[Jaybee00]

So they’ve completely missed the point even though it’s right in front of them. Inclusion criteria are far more important than PROMs.

Exactly.

How could you include patients with LC lung damage together with ME/CFS patients and expect to get any coherent results?

 

[Jaybee00]

Thread from one of the authors.

 

[poetinsf]

The patient criteria in the intervention studies listed in this article look vaguely like ME/CFS, except that they are much worse than 1994 Fukuda definition. It's like we are going back in time. And therein lies my criticism of LC studies: it's an amalgam of everything left over from covid as if PVFS is something new. It's no wonder they have problem figuring out what to measure. They really should separate ME/CFS subgroup of LC and then shovel money into researching it with ME/CFS standards rather defining new ad hoc criteria. ME/CFS phenotype, after all, is the most disabling type.

 

[Yann04]

Chloé de Canson writes some absolutely great philosophy related blog pieces and posts related Disability, ME/CFS, Post-COVID etc. It’s been a pleasure reading their blog.

I’m afraid they’re largely missing the point with relation to trial methodology here, though. They get the problem of clinical trials on an umbrella disease right but suggesting more PROMS is really not the solution at all and instead might contribute to further the problem.

 

[hotblack]

Thread from one of the authors.

That’s interesting, so it looks as if they have ME/CFS following LC, and are trying to work out how to study it. The recognition that PEM is key for ME/CFS and it is often incorrectly conflsted with fatigue is good too. As is the highlighting of FUNCAP.

I agree with @Utsikt though that there’s probably some things back to front here. Looking at it through a lens of PROMS rather than selection seems flawed. I can see their motivation though.