Tuesday, July 8, 2008

Atheroma Regression 101 -- Focus on HDL

Dr. Steve Nissen certainly deserves some respect. Not only does (a) he stick his neck out for the heart protection of Americans and (b) endorses CAC scoring/EBT for plaque/atheroma validation, he supports regression data in alignment with the Track Your Plaque goals for regression 60-60-60-60 already established by Dr. William Davis MD FACC. In a recent post-hoc analysis he reviewed data pooled from 4 major trials (ASTEROID, CAMELOT, REVERSAL, ACTIVATE) where plaque was diagnosed with angiography and tracked via IVUS over ~18 months (n=1455). The plaque volume % (PAV) decrements tracked well with i-n-c-r-e-a-s-i-n-g HDLs. And even more substantial reductions in total atheroma volume (TAV) were seen with more dramatic increases in HDLs.

The average 'regression' lead to an average 0.5 increase in PAV (2% increase and wide SD=15.7%) and a total atheroma volume reduction of 2.4 mm3 (only -0.2%; wide SD=17.4%). The authors note that in the ACTIVATE trial nearly 90% of patients were already on a statin and in fact their LDLs were less than 100 at baseline. Benefits for plaque stabilization and regression may have already been reaped prior to initiation for this sub-group which makes up 1/4 of the study. The duration of statin therapy was unknown.

These changes (avg plaque change: + 0.5)) were observed with the combination:
(1) LDL reduction (study avg=87.5 mg/dl final)
(2) increased HDL changes (the 'higher the better' with avg change=7.5% --the greatest regression was associated with HDL increases of 40% and HIGHER)
(3) lower LDL/HDL ratio (study ratio=2.1)

Average apo B was also significantly reduced from 131 mg/dl to 95 mg/dl (36% reduction) which was statistically correlated to changes in both PAV and total atheroma volume (p less than 0.001). The regression graph (see Figure and graph areas where PAV change is less than zero) lines up with TYP standard goals 60-60 for LDL and HDL.

TYP Goals (standard):
-LDL-C =60 mg/ml (see above arrow; higher?? see hard cardiac events)
-HDL-C = 40% increase to 60 mg/ml (study baseline avg = ~43 mg/dl; sorry--arrow wrong!)
-LDL/HDL ratio = 60/60 = 1.0

This regression data certainly emulates TYP to me!! The results truly seem to demonstrate the '60/60' vision...of a regressionist!! Like our beloved Dr. Davis :)

Often Dr. D achieves lipoproteins even beyond his standard goals. Many who follow TYP obtain LDLs lower than 60 and HDLs far higher than 60. Of course TGs fall far lower than 60 as well (often 40s) and of course apo B naturally drop as well (I'd conjecture less than 60).

So perhaps standard goals may be revised for those with 'stubborn' plaque.... based on this new information.

Who might benefit? The analysis showed that independent predictors of change in PAV were baseline PAV, the presence of diabetes and/or hypertension. For total atheroma volume, independent predictors of change were baseline TAV and BMI.

Refer to the L-sided endpoints of each graph in the Figure. The regression data certainly supports substantial size reductions of plaque occur with further improvements in LDL and HDL. HHHhhhhmmm..... Therefore, based on reported values of TYP'ers who experience regression/stabilization, the below standards may apply. At TYP, we are not LDL-centric. Small-LDLp centric extremely so, but never accuse us of LDL-mania...

Ultimate TYP (hypothetical) for Ultimate Regression:

-LDL-C = 60 mg/dl
-HDL-C = 60-80s mg/dl
-TG = 40-50 mg/dl
-TG/HDL ratio = ~50/83 = 48/80 = ~40/67 = 0.60
-LDL/HDL ratio = 60/80 = 0.75

-(? 60-60-60-0.60 ===> LDL-25(OH)D-apoB-TG/HDL-ratio)

The most significant contribution that Nissen made I believe was actually not mentioned in his comments or conclusions. In fact his conclusions from the data differed from mine regarding reduction of clinical events at 18 months and less than 18months after starting a Statin (or for ACTIVATE subjects, statin duration prior to the study longer than the rest of the test subjects). If you review Dr. Nissen's Table 6 below, you'll notice an interesting phenomenon. The best reduction of clinical events outcomes (obstructed events or events requiring re-vasc) supports actually a higher LDL-C (calculated) v. lower LDL-C at time 18 months for the group with a 'good' healthy HDL increase and LDL surprisingly g-r-e-a-t-e-r than average LDL=87.5 mg/dl.... hhhmmm... what could be going on here...? May stabilization (which translates best to reduced revasculariazations) require a little increase in % atheroma volume (though total plaque mass show decreases)?? The data may suggest this (see below 2 duplicated Table 6's).

Dividing out the study pop into 4 subgroups compared responses related to HDLs (and apo B) which appear to exhibit the highest association with clinical events. Perhaps when LDL is less than 87.5 mg/dl, the presence of small LDL still plays a HUGE role? Perhaps those individuals who experienced the most substantial INITIAL improvement in HDLs also had the greatest shift from Pattern B to Pattern A? Perhaps this shift explains the reduction in revascularizations (p=0.07) since these occur at higher frequency 6-12 months after a previous PTCA/stent. Rates of prior percutaneous intervention were in fact high in the study groups (71%, 30%, 81%, n/a). Could the first sub-group have had some plaque shrinkage but were still in the early stages of plaque-remodelling? Perhaps the first sub-group contained two distinct stages of regression -- one early (not stabilizied yet) and one late (stabilization + regression). Of course, the unstabilized will increase clinical events, especially if the presence of small LDL has not been addressed (ie, fish oil deficiency, vitamin D3 deficiency, carb-overloading). It only takes a few events or revascularizations to affect the data. Maybe there is a period of even 'fluffiness' of plaque (surface, core/internal) right before it stabilizes and regresses... synonymous with the changes occuring in the arteries where small LDL is shifting toward buoyant, more benevolent particles? Perhaps it's a phase... and we only have part of the snapshot...

What an amazing picture to behold though... Regression t-r-a-c-k-e-d in such a large population!

Let's thinc again... Perhaps the LDL-calculated based conclusions are bunk? Is LDL-calcuated really the 'lousy' cholesterol???

It's too bad the large trials failed to VAP or NMR the lipoproteins (or at least failed to report smLDL-particles).

Nicholls SJ, Tuzcu EM, Sipahi I, Grasso AW, Schoenhagen P, Hu T, Wolski K, Crowe T, Desai MY, Hazen SL, Kapadia SR, Nissen SE. (Full PDF here)
Statins, high-density lipoprotein cholesterol, and regression of coronary atherosclerosis.
JAMA. 2007 Feb 7;297(5):499-508.
PMID: 17284700

The conclusions that I draw from the data (though apparently non-statistically significant; who the heck was the statistician) is that human clinical events data support an approximately 50%+ decrease in revascularization-events in the group with higher LDL-C (read: lower small-LDLp) when compared with those with lower LDL-C (read: still high small-LDL in a subset of patients which may have contributed to higher clinical events requiring revascularization).

When plaque stabilizes, the echogenicity increases on ultrasound. Does this also contribute to mildly increased plaque volume? Perhaps? When you get a scab on a traumatic wound, does the scab appear thicker and heavier (esp if you're swimming!) before subsequently falling off after new pink epithelium grows?

Dr. Nissen notes:
"Greater percentage increases in HDL and lower levels of LDL (NOT) and LDL/HDL ratio in patients during treatment with a statin resulted in atheroma regression (BUT NOT HUMAN CLINICAL CAD/CVD EVENTS)...

...This is, to our knowledge the first time that increases in HDL levels have been shown to be an independent predictor of beneficial outcome with statin therapy."

Now Nissen is erudite.

He also looked at 'substantial' responders, meaning those with massive regression (b/c the above data otherwise makes little sense -- to statinators). And HDL-cholesterol undoubtedly must be raised for regression. Could Crestor have higher affinity for PPAR receptors or indirectly affect PPAR more?

His results when looking from a different perspective comparing Non-Responders v. Responders is even more intriguing than the above data. And holds more answers to the key to SUBSTANTIAL REGRESSION. Additionally this data is in direct compliance with TYP principles 60-60-60. Responders were assigned if significant volume/PAV% relative reduction of 5% or greater regression occurred (?about twice or more the average amount?).

Responders comprised 34% of the study population (n=370 were they all Crestor/ASTEROID??). Looking at this data (see Table 5 in the PDF), you'll notice immediately that huge increases in HDL and apo B occurred and likely lead to the dramatic PAV reductions in responders. In fact the difference between HDL improvements between Non-regression subjects and Substantial-Regression subjects was nearly 58% (6.5% v. 10.3%; wide SD 17.2%)) Remember average HDL change was 7.5%. Wow. I bet some of these individuals with massive plaque regression hit the TYP goals of 40% increase to HDL=60 mg/dl (perhaps combined with exercise/diet changes) and TG=60 mg/dl. For apo B, there was also wide variation apparently among the subjects, but also apo B dropped substantially by 40.9 mg/dl to an average of 90.1 mg/dl in the Regression group (almost close to TYP less than 60-70 mg/dl goals!). The median change was 34.8% (SD wide 25.0%) whereas the Non-regression median apoB change was only 25.7% (SD 24.8%).

Wouldn't it have been interesting to see if the Responders had reduced clinical events? Instead of just breaking down LDL-subsets?

Does Nissen know that Statins substantially affect PPAR???! And the regression benefits for regression may be attributed more strongly to PPAR than LDL? Evidently not... though he mentions VA-HIT and Helsinki Heart (he forgot DAIS) which are primary and secondary prevention trials that demonstrated reduced mortality and events with fibrates (PPAR-alpha agonists) especially in the diabetes subsets (ie, high insulin).

We discussed how statins both bind and indirectly affect plaque-mac PPAR earlier here.

The many pleiotropic benefits of statins are related to reduction in inflammation via Macrophage's PPAR activation at the plaque level in damaged endothelium:
--reduction in CRP
--plaque stabilization
--reduced oxLDL in plaque
--atheroma regression

Does this suggest that PPARs regress plaque...?

I do believe so

How do statins causes lipoprotein benefits via activation of Mac PPAR a and PPAR g ?

Nissen saw no significant contribution of statins on TGs on changes in PAV but TGs were statistically associated with improvement. The regression was attributed most strongly to vast HDL increases he reports (and to dramatic apoB and small-LDLp reductions which he didn't really address in the text) .

PPAR activation leads to these lipoprotein benefits:
--reduction in small-LDLp
--reduction in apoB
--reduction in TGs

So statins are like PPAR agonists? So statins work like food and food sensors?

What food item works the best as a PPAR agonist? What's my favorite supplement?

Bays et al nicely describes the benefits, MOA, and clinical implications of omega-3 fatty acids. Next to the NIACIN-king... fish oil R-O-C-K-S. Fish oil raises HDLs by an impressive 11-14% in clinical trials. The benefits are dose related, the higher the dose, the higher the HDLs. The lower the baseline HDL, the higher the potential increases.

Fish oils benefits for regression are three-fold:
(a) reduced systemic inflammation via PPAR
(b) increase HDLs via PPAR
(c) reduce apoB and insulin via PPAR

Another dimension of fish oil CAD benefits is its electro-stabilizing effects on cardiac conduction and prevention of arrhythmias.

Bays HE, Tighe AP, Sadovsky R, Davidson MH.
Prescription omega-3 fatty acids and their lipid effects: physiologic mechanisms of action and clinical implications.
Expert Rev Cardiovasc Ther. 2008 Mar;6(3):391-409. Review. (Full
PDF here)
PMID: 18327998

Synergistically, Statins and Fish oil can work together to stabilize and regress plaque. The authors eloquently review how both have separate mechanisms which complement and accelerate the conversion of small-LDLp to IDL than to large non-atherogenic LDL.

Synergism... I like that...

What are other mechanisms to raise and maximize HDL?

Depending on the degree of plaque, the more the below are achieved, the greater the increase in HDLs...
--stop all wheat
--stop all grains
--minimize all fruit
--stop all HFCS garbage
--consume adequate high-quality protein 1 g/kg
--consume Hg-free seafood
--consume EPA+DHA
--consume nuts/seeds
--intermittently fast
--exercise, some HIT (high intensity training -- only after medical clearance -- of course)
--get lean muscle mass
--abdominal fat loss
--decrease IR (see above)
--tobacco cessation
--maintain good sleep



Anonymous said...

Hey G,

Has anyone told you that you are incredible lately?


Dr. B G said...


Yes...my subliminal mirror, mirror on the wall. Remember, SNL? I love you too...


Stephan Guyenet said...

Hi G,

I just read your comment on Dr. Davis' blog about grains deranging vitamin D conversion. Do you have references for that? I'd be interested in checking them out. Thanks!

Dr. B G said...

Peter does the topic better justice than I.


I think I've seen also some references to the effect that leaky gut from celiac/silent-celiac reduces GI absorption of fat-soluble vitamins incl Vitamin D (if consuming foods like cod liver oil, liver, lard, etc).

Hope that is helpful!

Stephan Guyenet said...

Thanks G. Celiac definitely impacts the absorption of fat-soluble vitamins. There was also that study I came across a while back showing that wheat bran increases serum D turnover... Just wondering if I could get some mechanism. I guess there's not much money in bashing wheat.

Dr. B G said...

I'd love to see that mechanism as well! The great majority of gastric bypass patients become vitamin ADEK deficient as well. I wonder if the wheat and grains they eat exacerbate their already compromised vitamin/nutritional states?

. said...

Low HDL Cholesterol is Associated with Lower Gray Matter Volume in Cognitively Healthy Adults. - http://www.ncbi.nlm.nih.gov/pubmed/20725527

Dr. B G said...

Hi Ricardo,

Haven't seen that but it makes perfect sense! HDLs are may be a satellite extension of the neurologic system, besides being associated with the same inflammatory disorders that are associated with dementia and cognitive decline.

Chris K. just posted this:

Higher lipids, better health.