Thread by @ZahcM

Decided to make a thread on the failures of apoB or LDL-P. All related info on this topic will go here.

The AURORA trial: no relation between LDL or ApoB and major cardiovascular events - karger.com/Article/Abstra…

"Importance of LDL/HDL cholesterol ratio as a predictor for coronary heart disease events in patients with heterozygous familial hypercholesterolaemia: a 15-year follow-up (1987-2002)" - ncbi.nlm.nih.gov/pubmed/12755140

Looking at the supplement, baseline LDL particle number was not a better predictor than the standard TC/HDL ratio. It was concluded that particle composition may be more important than particle number - sciencedirect.com/science/articl…

"We observed that the association of apoB 100 with CHD was more attenuated by lipid and nonlipid risk factors than was LDL-C . . . HDL-C appeared to be the primary lipid predictor" - ncbi.nlm.nih.gov/pubmed/15492318

ApoB not independently predictive of risk after standard risk factors and the ratio of total to HDL cholesterol were considered -ncbi.nlm.nih.gov/pubmed/2062328

TC/HDL ratio better than apoB - ncbi.nlm.nih.gov/pubmed/15277429

This suggests that once the TC/HDL ratio is low, higher apoB doesn't affect atherosclerosis - ncbi.nlm.nih.gov/pubmed/27392507

LDL-P not associated with risk after adjustments for lipid + non-lipid factors - ncbi.nlm.nih.gov/pubmed/25477346

Ravnskov to Sniderman et al. in 2003 - ncbi.nlm.nih.gov/pubmed/12801767

No relation between apoB and events in PROSPER - ncbi.nlm.nih.gov/pubmed/16275871

TC/HDL stronger predictor than apoB (greater HR). ApoB added nothing in models - ncbi.nlm.nih.gov/pmc/articles/P…

ApoB no relation with all-cause mortality and CV events - ncbi.nlm.nih.gov/pubmed/26370413

LDL-P not related with CVD events - ncbi.nlm.nih.gov/pubmed/28728064

The TC/HDL ratio better than apoB in all analyses in FIELD - ncbi.nlm.nih.gov/pubmed/20526762

The TC/HDL ratio better than apoB in TNT/IDEAL analysis - ncbi.nlm.nih.gov/pubmed/18519851

AFCAPS analysis often cited for use of apoB but there was no association for change in apoB and events. On the other hand change in ApoA1 was - ncbi.nlm.nih.gov/pubmed/10662743

Claimed relation for apoB and CV mortality, but unconvincing study due to relation of lipoproteins with metabolic syndrome and insulin resistance in this population... and no relation with all-cause mortality -academic.oup.com/eurheartj/arti…

In this population, apoB was not associated with CVD and inversely associated with all-cause death - europepmc.org/articles/pmc58…

ApoB no independent association with IMT. OxLDL also remained associated after adjustment for apoB - ncbi.nlm.nih.gov/pubmed/15205217

One of the earliest studies to test apoB on disease (angiography) in comparison to other lipid markers - ncbi.nlm.nih.gov/pubmed/6651909

HDL and ApoA1 better discriminators of disease than apoB. For severity of disease apoB showed weak correlation (<0.3); overall ratios clearly best.

1985 study attempting to correlate apoB with disease, overall failure - ncbi.nlm.nih.gov/pubmed/3923814

2018 study from China concluding that small-dense LDL and OxLDL appear better than apoB for predicting CAD - ncbi.nlm.nih.gov/pubmed/29730953

ApoB was not associated with CAD, OxLDL was - "This study supports the hypothesis that lipid oxidation plays a role in the development of atherosclerosis" - ncbi.nlm.nih.gov/pubmed/11254911

On the other hand, this study showed no relation between changes in OxLDL and carotid IMT, but apoB changes showed no relation either - ncbi.nlm.nih.gov/pmc/articles/P…

Baseline, follow-up, or changes in ApoB not related to changes in PAV, TAV, and other measures of plaque composition (see supplement of paper) - ncbi.nlm.nih.gov/pubmed/28679524

LDL or ApoB had no relation to plaque regression; A "lifestyle modification score" was associated with plaque regression which was also correlated with increases in HDL-C - ncbi.nlm.nih.gov/pubmed/20234098

The ApoB/ApoA1 ratio was the strongest predictor of CV mortality in this study, but buried among the over 3000 word paper was the most important result in few words: “ApoB had no association with mortality from all causes” - ncbi.nlm.nih.gov/pubmed/16525840

ApoB not related to CHD events in women - ncbi.nlm.nih.gov/pubmed/8546747

Later report on all-cause mortality (ACM) found "no significant prediction of ACM by various lipid and lipoprotein parameters" - ncbi.nlm.nih.gov/pubmed/19849748

In PREVEND apoB was clearly not better than the TC/HDL ratio and worse than the apoB/apoA1 ratio and non-HDL -ncbi.nlm.nih.gov/pubmed/21129046

ApoA1 best predictor in this elderly population, not apoB - ncbi.nlm.nih.gov/pubmed/17234819

ApoB failed to predict mortality - ncbi.nlm.nih.gov/pubmed/22693160

Placebo group of HPS: LDL-P and apoB not better than LDL for CHD and completely failed to predict ischemic strokes. Also, poor old non-causal HDL-C remained significantly inversely related with CHD even after LDL-P adjustment - ncbi.nlm.nih.gov/pubmed/22539783

TC/HDL ratio stronger predictor than apoB in this study (greater risk ratios and if anything greater ROC). The figure below shows no increased risk for apoB in the lowest TC/HDL (full credit to @FatEmperor for this one) - ncbi.nlm.nih.gov/pubmed/17698856

Framingham analysis: apoB no better than the TC/HDL ratio - ncbi.nlm.nih.gov/pubmed/17699011

Note: for apoB to be a relevant predictor it needs to be proven that it is better than the TC/HDL ratio. Majority of studies show that it is not better or actually worse.

Analyses in LIPID generally showed better predictive value for the TC/HDL ratio than apoB - ncbi.nlm.nih.gov/pubmed/11889008

Similar was seen in a later analysis - ncbi.nlm.nih.gov/pmc/articles/P…

apoB not related to raised lesions; non-HDL + HDL added more to the basic model than apolipoproteins. Clear that lipids or apolipoproteins can barely explain the variations seen in atherosclerosis (low R2 values) - ncbi.nlm.nih.gov/pubmed/10073983

ApoB not related to vascular or nonvascular death - ncbi.nlm.nih.gov/pubmed/9261250

ApoB not associated with increased atherothrombotic events or mortality in CORONA, even associated with better prognosis; ApoA1 best predictor - ncbi.nlm.nih.gov/pmc/articles/P…

ApoB couldn't even survive a simple cholesterol adjustment in this one - ncbi.nlm.nih.gov/pubmed/11114956

EPIC Norfolk: risk for LDL-P attenuated after adjustment for HDL and triglycerides becoming no better than LDL-C - ncbi.nlm.nih.gov/pubmed/17276177

LDL-P fails highlighted in the image below. Also: "For none of the lipoproteins was change from baseline to year 1 significantly associated with CHD" - ncbi.nlm.nih.gov/pubmed/18599797

ApoB added nothing to simple total cholesterol - ncbi.nlm.nih.gov/pubmed/1585854

ARIC: apoB not better than the non-HDL/HDL ratio (HRs actually greater for the lipid ratio). ApoB even lost its relation to CHD in models adjusted for non-HDL-C and HDL-C in obese subjects - ncbi.nlm.nih.gov/pubmed/23109406

In an early report of the Women's Health Study, after exclusion of very soft endpoints, apoB was clearly a weaker predictor than the TC/HDL ratio, and even far inferior to CRP - jamanetwork.com/journals/jama/…

Small-dense LDL and apoA1 related to CV events whereas apoB lost significance in multivariate models - ncbi.nlm.nih.gov/pubmed/29512264

In those without diabetes or metabolic syndrome, neither LDL nor LDL-P predicted CV events after adjustments - sciencedirect.com/science/articl…

New large study: apoB and apoA1 does not improve CV risk prediction beyond total cholesterol and HDL cholesterol - ahajournals.org/doi/10.1161/CI…

Very interesting data in the supplement for fatal CVD where HDL and ApoA1 were the strongest (inverse) predictors.

From BioBank study above:

apoB no better than HDL for CV events.
Assoc. for apoB weaker in some groups (women, elderly, lower triglycerides i.e. apoB relations depend on other factors).
Low cholesterol ⬆️ CV deaths.
Low HDL or ApoA1 bigger risk factors than apoB for CV death.

Only became aware of this from @FatEmperor and @TuckerGoodrich.

The Kitavans, who appear absent of CVD, have apoB levels generally on the higher end well-above recommended levels (<90 mg/dl).