Thread by @DaveKeto

1/ Really great thread by @Lpa_Doc in response to @ethanjweiss.

Read completely before returning here for my added thoughts...

[Warning: this one will get a little geeky]

https://twitter.com/Lpa_Doc/status/1368245738908512258

2/ Per @Lpa_Doc's (2) and (3) in the thread, when we have "normal" TG levels (guidelines would say < 150, I prefer < 100) then most cholesterol in ApoB-containing lipoproteins (ApoB-Lp) are in LDL and there's little VLDL-cholesterol...

3/ But where TG is high, there's more utility in looking at ApoB given it lumps remnants (VLDL/IDL) with LDL.

-- But IMO, that's also the problem. We want to disentangle where remnant association is much more relevant than LDL association...

4/ I don't typically like ratios, but if I'm getting arm-twisted, here are the two I'd lean toward:

1) OxPL-ApoB / ApoB

or

2) Non-HDL / ApoB

That second one isn't highly studied, but it should be. We already have a lot of metrics on it in existing datasets everywhere...

5/ And it's relevant to this topic with regard to ApoB-Lp cholesterol on a per-particle basis.

NHANES data shows a high Non-HDL / ApoB to be associated with lower all cause mortality (but that was back-of-the-envelope analysis, will go deeper with biostat help eventually)

6/ So in short, I don't like looking at ApoB in isolation as I think it's a safe bet it will be much less predictive than looking at it combination with either OxPL-ApoB or non-HDL.

(Regardless, we're getting all this data and more in the LMHR study, of course.)