REMAP-CAP has dropped insanely underpowered in terms of its mortality endpoint (117 patients in one arm) based on statistical power & methodologic problems, it adds nothing to CAPE-COD my take hasnt changed since my discussion of CAPE-COD here https://emcrit.org/pulmcrit/cape-cod/ #EMIMCC - ThreadSky

pulmcrit.bsky.social • 16 days ago

REMAP-CAP has dropped

insanely underpowered in terms of its mortality endpoint (117 patients in one arm)

based on statistical power & methodologic problems, it adds nothing to CAPE-COD

my take hasnt changed since my discussion of CAPE-COD here https://emcrit.org/pulmcrit/cape-cod/ #EMIMCC

Reposted from Antibiotic Steward 🅱️C🆔🅿️🌟

🔥REMAP-CAP RCT🔥
Among patients admitted to the ICU with severe non-pandemic CAP & either cardiovascular or respiratory failure there was a low probability that a 7d fixed-duration course of hydrocortisone improved 90d mortality @pulmcrit.bsky.social #EMIMCC #IDSky
link.springer.com/article/10.1...

Comments

absteward.bsky.social•15 days ago

Thanks, are you going to update the page adding REMAP-CAP? #IDSky #EMIMCC

pulmcrit.bsky.social•15 days ago

not planning to, I usually don’t update the pulmcrit blogs unless they’re really wrong. I do update the IBCC chapters.

almoskow.bsky.social•15 days ago

A cautionary tale about highly complex trial designs? Difficult to interpret. Small groups powered into for large effect sizes. Risk of error?

emlitofnote.bsky.social•15 days ago

So not 🚨🔥🚨🔥🚨🔥🚨 is what you're trying to say

pulmcrit.bsky.social•15 days ago

hype >>>> evidence.

I’ve been seeing messaging that this trial was going to overturn the results of CAPE-COD and that’s just not the case.

but I do admit (and have consistently said for a long time) that any mortality benefit (if real) is super small and very hard to replicate

tomboylesid.bsky.social•15 days ago

For me it's a bit like statins. They do work, but the effect sizes are very small and so people will endlessly bicker about them. Personally I tend to lose interest when there are interventions with large effect sizes to consider.

tomayates.bsky.social•14 days ago

Clearly, in a resource constrained system, it is critical to ensure high coverage of high impact interventions. But do you not think that, in a lot of conditions, improvements come about as a result of multiple small wins? High impact single interventions, e.g. ART, are unusual.

tomboylesid.bsky.social•14 days ago

It's about bang for your buck. When money and human resources are limited and you haven't got good coverage of high impact interventions, then that is where the focus should be. In high resource setting, things shift and incremental gains become worth it.

tomboylesid.bsky.social•14 days ago

I'm looking for a NNT of about 50 to save a life for most interventions, obviously depending on how much resources they consume. This would be very different if I worked in a high resource setting. E.G. until we have better ART coverage and retention, statins for HIV patients are not on the radar.

emlitofnote.bsky.social•15 days ago

The clear answer is multiple small trials with threats to internal validity and heterogeneity between protocols and populations, followed by a meta-analysis manifesting the authors' pre-conceived notion regarding its utility.

pulmcrit.bsky.social•15 days ago

then everyone gets published and promoted 🙌

tai-md.com•15 days ago

What’s the take?

pulmcrit.bsky.social•15 days ago

details here https://emcrit.org/pulmcrit/cape-cod/

I think steroids work, but there isnt a reproducible/robust mortality benefit

everyone keeps having a screaming match about whether steroids cause a massive mortality difference & I think that’s the wrong question

joealcock.bsky.social•14 days ago

When there is uncertainty, it doesn’t help to only halt trials early if they trend towards showing harm or null results.

Comments

Posting Rules

Reply