"The LDL data is a complete reversal of what we expected, if that's fair to say?" stated Cancundude.

There was never any anticipated difference in apabetalone benefit for above- or below-median LDL-C groups. All patients had well controlled LDL-C with median of ~65 mg/dL at baseline that increased about 12-15% at 100 weeks vs. baseline in both groups. In other words, apabetalone treatment had no apparent effect on LDL-C. Howevever, most folks don't appreciate the heterogeneity of apolipoprotein B-containing lipoprotein metabolism. 

My own hunch is that this has something to do with the abnormal lipoprotein profiles associated with insulin resistance and diabetes. Not all LDL particles are created equal. LDL particles and the cholesterol they carry (LDL-C) are a product of both apo-B containing lipoprotein production and catabolism. The liver hepatocyes synthesize and secrete into the circulation the nascent apo-B containing lipoprotein as a large particle filled with a lot of triglyceride (TG) and cholesterol ester (CE) in its core. These are called very low density lipoproteins, or VLDL. In diabetes and insulin resistance, VLDL are even more TG-rich than the VLDLs seceted by healthy insulin sensitive folks. When metabolized/catabolized, peripheral tissues take up much of this TG core for energy leaving behind in the circulation a smaller particle, low-density lipoprotein (LDL) with proportionally less TG and more CE.

The end result is not just one uniform LDL particle size and not one uniform core ratio of TG:CE. There is a wide spectrum of LDL particle sizes and lipid compositions. The most atherogenic of these seem to be the small, dense LDL particles. There are two well established patterns of LDL particles: Pattern A and Pattern B. To copy from the ref below: 

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5441126/

"Numerous clinical studies have been conducted to establish the link between the composition of circulating LDL particles and the risk of atherosclerosis and CVD development. According to the current consensus, 2 main phenotypes, A and B, are defined based on plasma LDL profile, with intermediate A/B phenotype lying in between [11]. The phenotype A is characterized by the predominance of large buoyant LDL (lbLDL) and the phenotype B by the predominance of sdLDL [12, 13]. Phenotype B was reported in a number of diseases, including metabolic disorders [14, 15], obesity [16, 17], and type 2 diabetes [18, 19] and is considered as a risk factor of coronary heart disease (CHD)."

A couple other refs on differences between Phenotype A and B particles, including lipid composition changes:

https://www.ahajournals.org/doi/full/10.1161/01.atv.16.8.1040

https://www.ncbi.nlm.nih.gov/pubmed/12200749

Phenotype A is mainly large and intermediate LDL. Phenotype B is mainly small and very small LDL. Keep in mind clinical measurement of LDL-C is just a sum of all LDL-C in the blood. Furthermore, there is one apoB molecule per LDL particle. So for a given LDL-C value in the blood, a person with Phenotype B will have more LDL particles and total apoB than a person with Phenotype B because small dense LDL contain less cholesterol per LDL particle. These small dense LDL particles are more atherogenic.

One possibility is that the below median LDL-C population in BETonMACE has an increased prevalence of pattern B exhibitting a high prevalance of small, dense LDL particles that are more atherogenic. It is also possible that the below median LDL-C population in BETonMACE is actually enriched in this pattern B; pattern A would have more cholesterol rich LDL particles that could lead to higher circulating LDL-C values but a lower number of and less atherogenic LDL particles. The metabolic disturbances that give rise to small dense LDL, the susceptibility of the small dense LDL to further damage and modification that accentuates their atherogenicity, the ability of these small dense LDL to insult the cardiovascular system, and the ability of the body to defend, adapt and recover from small dense LDL-C insults may all potentially be improved by apabetalone. 

Also, if you look at slide 32, you can see that the below median LDL-C placebo group had a higher MACE risk than the above median LDL-C placebo group, consistent with my suggestion above that the below median LDL-C group may be enriched in those with more atherogenic (and less cholesterol rich) LDL particles.

Just a hunch.

BDAZ